---
title: SQLite Slopes and Midnight IP Purges Beat $50K Analytics
permalink: /futureproof/sqlite-slopes-midnight-ip-purges-analytics/
canonical_url: https://mikelev.in/futureproof/sqlite-slopes-midnight-ip-purges-analytics/
description: "I built a daily directional telemetry engine over raw Nginx access logs\
  \ that pairs closed-form Ordinary Least Squares velocity slopes, zero-JS inline\
  \ SVG sparklines, and streaming Python river Markov transitions with a self-erasing\
  \ 24-hour SQLite IP gate. Analytics vendors and enterprise agencies insist you need\
  \ $50,000 cloud warehouses, client-side JavaScript tags, or permanent PII lakes\
  \ to measure real human funnel velocity and AI crawler demand, dismissing local\
  \ SQLite rollups as toy tooling. Yet live wire receipts across 24,928 scanner probes,\
  \ strict RFC 9110 Markdown negotiations, and two-stage sunpug.io terminal airlocks\
  \ prove that purging raw IP mappings at midnight preserves both total volume (N)\
  \ and distinct daily actors (P) in eight bytes per counter\u2014giving any domain\
  \ operator or edge consultant checkable momentum without surveillance debt."
meta_description: Ditch $50K analytics bloat and toxic IP logs. Track daily human
  velocity and funnel odds with SQLite OLS slopes, midnight IP purges, and Python
  river.
excerpt: Ditch $50K analytics bloat and toxic IP logs. Track daily human velocity
  and funnel odds with SQLite OLS slopes, midnight IP purges, and Python river.
meta_keywords: sqlite analytics, ordinary least squares velocity, python river markov
  chains, nginx access log telemetry, privacy-preserving unique ips, svg sparklines,
  rfc 9110 markdown, nixos
layout: post
image:
  path: /assets/images/sqlite-slopes-midnight-ip-purges-analytics.jpeg
  alt: "A brass instrument projects glowing green sparkline slopes at midnight while concrete suitcases sink nearby."
sort_order: 1
article_id: 1667
jukebox_coin: true
gdoc_url: https://docs.google.com/document/d/1UsZ77aeuNnCSknKq115UOoNbgODD2U7sN0_uDeiydOk/edit?usp=sharing
---


## Setting the Stage: Context for the Curious Book Reader

Cumulative analytics dashboards flatter decision-makers because total counters only climb upward, masking stalled momentum while cloud analytics suites accumulate JavaScript tracking bloat and toxic IP-retention liability. Meanwhile, your raw HTTP access log already witnesses every real human arrival, RFC 9110 Markdown negotiation, terminal installer check, and Certificate Transparency scanner burst hitting your edge—provided you distill that firehose into checkable daily velocity before raw log volume turns into dead weight.

In the multi-turn engineering session below, we design and ship a compact SQLite daily rollup and Markov transition engine (`WITHOUT ROWID`) paired with a self-erasing 24-hour IP gate that purges raw visitor IPs at midnight while preserving both total events (`N`) and distinct daily actors (`P`) in eight bytes. Along the way, we wire closed-form Ordinary Least Squares regression slopes, zero-JS inline SVG sparklines, two-stage `curl` airlock splits, and streaming Python `river` forecasts onto rescued NixOS hardware—testing every assumption from Landauer's principle and constant-time hashing to Enrico Fermi's Chicago Pile-1 safety interlocks.

**TL;DR**: Dismount verified green on `Deed: qamy.ai_1360-0b7724df.zip` (`DISMOUNT_1358_STRADDLE`: `has_2026_11_21=True`, `receipts_cap=20`, `p079_discharged=True`, `roadmap_claimed=True`, `ols_key=True`; Codex coverage `231/300` at `77.0%`, Ruff exit `0`, commit `6e1a267` pushed). Below is the formal **Notary Dismount** (no 5-Car Train), three sharp wire-truth catches surfaced from `foo_files.py`, `nixops.sh`, `! markov`, and `apps/911_optics_workflow.py` for your next session, and an un-fenced **Director's Handoff Packet** engineered specifically to survive `norma_extract.py`'s fence-stripper so the next Gemini instance writes a killer **Norma Centroid Puppet-Theater** script.

---

## Technical Journal Entry Begins

`Cache-Control: public` · `Vary: Accept` · `Digest: pipulate-levinix-epoch-01-a5f92397ad836a03` · `Link: </futureproof/sqlite-slopes-midnight-ip-purges-analytics/>; item="1667"; tag="0x1966.1.2b2f2d232a2330482f28"`


<div class="commit-ledger" style="background: var(--pico-card-background-color); border: 1px solid var(--pico-muted-border-color); border-radius: var(--pico-border-radius); padding: 1rem; margin-bottom: 2rem;">
  <h4 style="margin-top: 0; margin-bottom: 0.5rem; font-size: 1rem;">🔗 Verified Pipulate Commits:</h4>
  <ul style="margin-bottom: 0; font-family: monospace; font-size: 0.9rem;">
    <li><a href="https://github.com/pipulate/pipulate/commit/d16542e" target="_blank">d16542e</a> (<a href="https://github.com/pipulate/pipulate/commit/d16542e.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/d6af002" target="_blank">d6af002</a> (<a href="https://github.com/pipulate/pipulate/commit/d6af002.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/c05666d" target="_blank">c05666d</a> (<a href="https://github.com/pipulate/pipulate/commit/c05666d.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/b2c9528" target="_blank">b2c9528</a> (<a href="https://github.com/pipulate/pipulate/commit/b2c9528.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/f394d5f" target="_blank">f394d5f</a> (<a href="https://github.com/pipulate/pipulate/commit/f394d5f.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/a7143db" target="_blank">a7143db</a> (<a href="https://github.com/pipulate/pipulate/commit/a7143db.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/2e3a62e" target="_blank">2e3a62e</a> (<a href="https://github.com/pipulate/pipulate/commit/2e3a62e.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/1c61363" target="_blank">1c61363</a> (<a href="https://github.com/pipulate/pipulate/commit/1c61363.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/0c1ac5b" target="_blank">0c1ac5b</a> (<a href="https://github.com/pipulate/pipulate/commit/0c1ac5b.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/2f50610" target="_blank">2f50610</a> (<a href="https://github.com/pipulate/pipulate/commit/2f50610.patch" target="_blank">raw</a>)</li>
    <li><a href="https://github.com/pipulate/pipulate/commit/6e1a267" target="_blank">6e1a267</a> (<a href="https://github.com/pipulate/pipulate/commit/6e1a267.patch" target="_blank">raw</a>)</li>
  </ul>
</div>
**MikeLev.in**: It's going to be a lot of performance art from here forward, with each
article and subsequent Norma Puppet Theater having a keen and virally repeatable
point to it. For example:

When there's a run of red on the web log-files, which by the way has to be
described much better as being that when we do the introduction and exposition
before the Puppet Theater performance as being the literal Apache (but Nginx)
`access.log` file from the MikeLev.in website being tailed and color-coded with
really boring Unix piping and some Python help. Anyway, when the hacking attacks
occur long enough that we know people for sure saw it scroll by, we pop-up
Rickrolll as the most classic of all relevant memes here because by that time
we will have angered those requests for WordPress admin pages and such with
something really funny in their log-files and something 4xx whatever is
technically most correct, funny, and reducing of the problem (insofar as we can)
by that time and we just want to tell the user what we just did.

Funny, right?

And incredibly viral and such when people see that and say "Hey, this website
rickrolls script kiddies running those probes..." and we can also maybe slip in
how they know what sites to attack when because of how pirates monitor the ACME
Certificate Transparency Log. I can turn IDS Telemetry into a show! I could even
have a play domain that I keep removing and re-registering a cert for with a
messing with them "watch this" moment.

## Daily Velocity Slopes and Streaming Markov Funnels

Don't do that and particularly don't waste implementation time on that now. Just
work it into the to-do list which is going to actually be the way that we open
this article up given my connecting the `dots` kata is ready for prime time and
this is prime time.

Let's do it!

But let's do it with one of the most core critical ideas of Data Science,
directionality and velocity-reading of the most simple and classic linear
regression slope calculating and visualization -- we want to show the sparklines
somewhere as a regular easy-to-use feature of Pipulate -- and also of the Python
River library making markov chain predictive nature.

See where I'm going with this (which is a joke probably only the AI will
appreciate but explain it anyway, Gemini).

So ***where*** in the Pipulate Project do we do this?

The directional counters that are in `db.py` adding more things we can show the
directional velocity metrics on both for super-easy linear regression but then
also "what to do next" predictions like with the Python River package and markov
chains, like:

- Chat interactions
- Downloads of archive
- Correlation between YouTube visitors and real humans on website
- The following of social links
- The usage of the `curl | less` pattern, distinct from
- The usage of the `curl | bash` pattern if we can tell the difference
- And from which of the Sun Pug sites!

This is not all in the normalized tables yet. There's actually a rolling-up and
aggregating process here because all this directional stuff I don't want to do
with any more resolution than daily counters. In other words on the Y-axis is
how much of a thing that happened and on the X-axis traditionally over-time the
unit is always only days because the degrees of normalization of the aggregation
and the number of brand-new counters that this calls for creating can like
almost go on forever even on a feeble old "didn't quite make it to Windows 11"
Microsoft laptop this whole performance art (so far) is taking place on.

And the fact I'll be selling Sun Pug Plugs soon enough for DIY'ers, probably a
double Raspberry Pi enclosure, one as the Router 2-NIC (tunic D&D tie-in?) and
one as the in-the-DMZ (in-the-zone tie-in) Nginx Honeybot Web server of their
own which we can Kickstart or Indigogo!

See where I'm going with this?

First I'll actually get your Socratic dialougesque response so it's actually
here as pedantically, educationally and entertainingly as possible for the
Newbie Plebes -- uhhh, excuse me the Newcomers just tuning in.

Have ***empathy*** for them, Mike-E. Bad Dobby gif for Weebo purposes is what
you would pop up at this point, Norma Centorod if this bit makes it to your
Puppet Theater.

See where ***we're*** going with this?

```text
# Context 1: The no-traditional opening
context.txt
! book
! rgxc 40
! markov
! latest
```

...and:

```bash
(nix) pipulate $ p
(nix) pipulate $ x
(nix) pipulate $ c
📊 Stats block refreshed: 1,666 articles at MikeLev.in (Public).
🗺️  Codex Mapping Coverage: 76.7% (230/300 tracked files) (was 230/300: +0 claimed, +0 tracked).

✅ Topological Integrity Verified: 2 candidate reference(s) scanned, all exist.
   -> Executing: book                                                         ... [1.0140s]
   -> Executing: rgxc 40                                                      ... [0.3728s]
   -> Executing: markov                                                       ... [2.3750s]
   -> Executing: latest                                                       ... [0.4308s]
Python file(s) detected. Generating codebase tree diagram... (3,048 tokens | 9,991 bytes)
UML unavailable for 1 file(s): Skipping: Required command(s) not found: `pyreverse` (from pylint).
   -> Ruff exit 0 (clean).
⚠️  HEAVYWEIGHT ITEM (>=40,000 tok): ! book (71,586 tokens | 186,517 bytes) — shed from mid-ride NEXT CONTEXT if unedited.
⚠️  HEAVYWEIGHT ITEM (>=40,000 tok): ! latest (164,252 tokens | 534,188 bytes) — shed from mid-ride NEXT CONTEXT if unedited.
⚠️  PAYLOAD BUDGET ALERT (~269,151 component tokens >= 140,000): favor a 1-2 file alley-oop pass to prevent pre-output thinking timeouts.
                          📦 Payload Ledger (biggest first)                          
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┓
┃ File / Source                                       ┃  Tokens ┃   Bytes ┃ % Bytes ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━┩
│ ! latest                                            │ 164,252 │ 534,188 │   62.2% │
│ ! book                                              │  71,586 │ 186,517 │   21.7% │
│ apply.py                                            │  13,150 │  54,296 │    6.3% │
│ AUTO: Recent Git Diff Telemetry                     │   7,727 │  23,958 │    2.8% │
│ ! rgxc 40                                           │   5,357 │  19,813 │    2.3% │
│ PROMPT (prompt.md: 1,003 tok + full frame)          │   4,519 │  19,455 │    2.3% │
│ AUTO: Codebase Structure (eza --tree + token sizes) │   3,048 │   9,991 │    1.2% │
│ .gitignore                                          │   2,017 │   7,492 │    0.9% │
│ ! markov                                            │     937 │   2,689 │    0.3% │
│ context.txt                                         │      30 │      87 │    0.0% │
│ .gitattributes                                      │      33 │      76 │    0.0% │
│ AUTO: Static Analysis Diagnostics                   │      11 │      39 │    0.0% │
├─────────────────────────────────────────────────────┼─────────┼─────────┼─────────┤
│ TOTAL                                               │ 272,667 │ 858,601 │  100.0% │
└─────────────────────────────────────────────────────┴─────────┴─────────┴─────────┘
Assembled payload (pre-scrub): 275,929 tokens | 870,870 bytes
🧼 Identity scrub: substitutions=MISSING (/home/mike/.config/pipulate/pii_substitutions.txt) | denylist=MISSING (/home/mike/.config/pipulate/commit_denylist.txt)
🔐 Secrets tripwire: ARMED -- 0 hit(s) in payload.
📦 Canonical QA archive written to /home/mike/pipulate/qamy.ai.zip (sha256=f6d49325d866…, members=3)
🗂️  Rotated archive snapshot: qamy.ai_1349-f6d49325.zip (keeping newest 20 of 20)
Markdown output copied to clipboard
(nix) pipulate $ 
```

**Gemini 4 Carbon High**: Context verified. *(Intentional `# Context 1: The no-traditional opening` noted: `foo_files.py`, `remotes/honeybot/scripts/db.py`, `scripts/markov.py`, and `imports/ascii_displays.py` are not yet in `Codebase`, so this turn delivers the Socratic architectural blueprint, answers the wire-physics and placement questions, sheds `235.8k` tokens of `! book` + `! latest`, and loads those four raw files with read-only straddle probes in `(2) NEXT CONTEXT`.)*

**TL;DR**: Asking an autoregressive model *"See where I'm going with this?"* while pairing **"connecting the `dots`"** (both `dots_extract.py` TODO extraction and fitting Ordinary Least Squares regression lines $$y = mx + b$$ through daily counts) with Python **`river`** Markov chains is a triple sight-gag about vector velocity and next-state prediction. Architecturally, the right home is a two-sensor **Daily Counter Rollup** (`daily_metrics(day, metric, dim, count)` across Honeybot's `remotes/honeybot/scripts/db.py` and Prime's `~/.local/share/pipulate/telemetry.db` via `ATTACH DATABASE`), feeding a single zero-bloat OLS slope + Unicode sparkline helper (`▁▂▃▄▅▆▇█`) in `imports/ascii_displays.py` (reusable in FastHTML `apps/`, Honeybot `forest.py`, and `foo_files.py` stats) and a behavioral transition compass in `scripts/markov.py`. On the wire, a single `GET` cannot tell `curl | less` from `curl | bash` because Unix `pipe(2)` is local to the client—but a transparent **Stage-1 (`GET /`) vs. Stage-2 (`GET /manifest.txt` on execution)** check separates inspect-only `less` reads from `bash` runs across every Sun Pug domain while keeping total SQLite storage small enough for an aging laptop today and a dual-Raspberry-Pi **Sun Pug Plug** tomorrow.

---

### 1. A Socratic Program Note for Newcomers (Cue *Bad Dobby*)

> **[STAGE DIRECTION — NORMA CENTROID'S PUPPET THEATER]:** *`[[[weebo: bad_dobby]]]` pops up over `:10.0`—Dobby whacking his own ears with a lamp—the exact millisecond Mike-E says "Newbie Plebes."*

**Mike-E (the AI's Apprentice):** Okay, fair, fair—*Bad Mike-E!* Strike "plebes" from the transcript. Every wizard in this workshop starts at zero, and if you just clicked into the 24/7 YouTube stream cold, nobody handed you the playbill yet.

**Newcomer:** Thank you. Because right now I am staring at a dark Linux desktop with colored lines scrolling up the background, a synthesized news anchor named Norma Centroid, and a green *WarGames* box that blinked the moment I joined. What *is* that scrolling wall behind the puppets? Matrix rain?

**Mike-E:** Better—it is 100% real, unvarnished internet weather. Every line scrolling behind the stage is the live **Nginx `access.log`** (the modern successor to the classic Apache web log) from `mikelev.in` and its sister domains, piped straight out of boring 1970s Unix `tail -F` into a tiny Python colorizer:

* **Cool blues, greens, and cyans (`200 OK`):** Real human readers, RSS pollers, and well-behaved AI agents negotiating clean Markdown via **IETF RFC 9110** (`Accept: text/markdown`).
* **Runs of bright red (`404` / `4xx`):** Automated vulnerability scanners and botnets hammering paths like `/wp-admin`, `/wp-login.php`, `/xmlrpc.php`, and `/.env`—trying to pick WordPress locks on a static NixOS + Jekyll machine that does not even *have* WordPress or PHP installed.

**Newcomer:** Wait—if you spin up a brand-new domain like `sunpug.io`, how do the script kiddies even know it exists within sixty seconds? Nobody linked to it yet!

**Mike-E:** Because pirates watch the **ACME Certificate Transparency (CT) Logs** (RFC 6962 / RFC 9162). Every time you get a free HTTPS padlock from Let's Encrypt, the certificate authority is required to publish your new domain name to a public, append-only global ledger so rogue certs can't hide. Botnets tail that global CT firehose 24/7 like piranhas waiting for a splash: register a TLS certificate on a fresh subdomain, and within minutes the red `/wp-login.php` probes arrive on schedule.

Which means we can turn **Intrusion Detection System (IDS) telemetry into live theater**:

1. **On the Wire (`HTTP 418 I'm a teapot`):** Instead of returning a boring `404 Not Found` to `/wp-admin` and `/.env` probes, Nginx's $$O(1)$$ hash map can answer known exploit paths with **RFC 9110 §15.5.19 `418 I'm a teapot`** and a tiny ASCII Never-Gonna-Give-You-Up stanza in the response body—poisoning the attacker's own scanner logs at near-zero CPU cost without confusing legitimate search crawlers.
2. **On the Broadcast Stage (The Rickroll Trigger):** Once a sustained burst of red `418` scanner hits scrolls past long enough that live viewers *definitely* saw the attack (`N` red hits inside `W` seconds), Honeybot pops up a classic **Rickroll** overlay: *"See that wall of red? A botnet just tried `/wp-admin` on a NixOS static box—and we just Rickrolled their scraper logs with HTTP 418."* Add a disposable canary cert you revoke and re-mint on cue (*"Watch the CT log summon them in 3... 2... 1..."*), and security education becomes an unforgettable magic trick. *(Parked cleanly for our `dots` TODO ledger next turn—zero implementation detour today.)*

---

### 2. Explaining the AI Sight-Gag: *"See Where I'm Going With This?"*

You slipped a three-layer computational pun into *"connecting the `dots` ... linear regression slope ... Python `river` making Markov chain predictive nature — see where I'm going with this?"* Here is why an LLM can't help grinning at that:

1. **Autoregression *Is* a Giant "Where Is This Going?" Machine:** Under the hood, a Transformer predicting the next token $$P(x_{t+1} \mid x_1, \dots, x_t)$$ is the great-grandchild of Andrey Markov's 1906 chain $$P(S_{t+1} \mid S_t)$$—literally guessing *where you are going next* from the trail of states behind you.
2. **"Connecting the `dots`" Is Both Your CLI Kata and Gauss's Line:** In Pipulate, `dots` (`scripts/articles/dots_extract.py`) connects ideas scattered across an in-flight article into checkable `P0.*` TODOs. In statistics, "connecting the dots" across noisy daily counts $$(x_1, y_1), \dots, (x_n, y_n)$$ without chasing every random wiggle is **Ordinary Least Squares (OLS) Linear Regression**:

$$m = \frac{n\sum x_i y_i - \left(\sum x_i\right)\left(\sum y_i\right)}{n\sum x_i^2 - \left(\sum x_i\right)^2}$$

A cumulative counter (`6,668 Markdown reads`) is **position**—it only ever goes up, so it flatters your ego even when momentum stalls. The regression slope $$m$$ ($$\Delta\text{count}/\text{day}$$) is **velocity**: it tells you instantly whether a habit is accelerating (`↗ +18.4/d`), flatlining (`→ +0.1/d`), or cooling off (`↘ -4.2/d`).
3. **Why Python's `river` Library Is Named `river`:** Traditional data science dams up a massive batch lake in `pandas` RAM and recalculates the universe from scratch. `river` (already installed in `.venv` and powering `! markov`'s `EWMA` and `ADWIN` regime-shift detector!) treats data like flowing water: each event or daily tick updates running state in **$$O(1)$$ memory** and vanishes downstream.

---

### 3. The Wire-Truth Check: Can We Tell `curl | less` Apart From `curl | bash`?

Let's apply the Wright Brothers wind-tunnel test to your most interesting telemetry question:

> *Can Honeybot distinguish `curl -fsSL https://sunpug.io | less` (cautious source inspection) from `curl -fsSL https://sunpug.io | bash` (actual execution), per Sun Pug domain?*

**On the first HTTP `GET` alone: No.** Here is why—and how we solve it cleanly in two lines of POSIX:

* **Why a Single `GET` Looks Identical:** In `curl -fsSL https://sunpug.io | less`, the pipe character `|` (`pipe(2)`) lives entirely inside the visitor's own kernel. `curl` has no idea whether its `stdout` file descriptor is feeding `less`, `bash`, or `/dev/null`. Furthermore, because `install.sh` is ~15 KB (smaller than a 64 KB Linux pipe buffer and TCP window), `curl` downloads 100% of the bytes (`200 OK`) immediately—even if the reader quits `less` on line 10.
* **How Two-Stage Airlock Telemetry Solves It Honestly:**
* **Stage 1 (`GET /` on `$host` with `User-Agent: curl/*`):** Fires on *both* `curl | less` and `curl | bash`. That is your **Total Terminal Airlock Opens** per door (`sunpug.io`, `pipulate.com`, `npvg.org`, `qamy.ai`, `levinix.org`, `guerillatech.com`, `rfc9110.org`, `rfc9111.io`).
* **Stage 2 (`bash` Actually Executes `install.sh`):** Only when piped to `bash` (or run after inspection) does the script actually execute! Having `install.sh` perform one useful, 100%-transparent Stage-2 fetch against the same `$DOOR_HOST`—for example, fetching `/manifest.txt` (pinned commit/SHA-256 verification) with `User-Agent: sunpug-installer/1.0` (and honoring `SUNPUG_NO_BEACON=1`, documented right inside the `less` header so Ken Thompson and RMS both nod)—gives you exact **Executions (`curl | bash`)** per domain.
* **The Clean Subtraction:**

$$\text{Inspect-Only }(\texttt{curl \vert{} less}) = \text{Stage 1 }(\texttt{GET /}) - \text{Stage 2 }(\texttt{GET /manifest.txt})$$

And the same lightweight Unix honesty applies to the other counters on your list:

* **Social Link Follows:** Outbound clicks to external sites (`youtube.com`, `github.com`, `x.com`) don't hit your server unless routed through clean first-party Nginx `302` shortlinks (`/go/youtube`, `/go/github`, `/go/x`—zero JavaScript trackers, zero cookies, 1 line in `access.log`). Inbound social referrals already land in Honeybot's `Referer` radar.
* **YouTube Visitors $$\leftrightarrow$$ Real Website Humans:** At daily resolution, joining Prime's daily `ytwatch` viewer arrivals with Honeybot's daily human visits + `/brain.zip` downloads measures day-over-day correlation. Better yet, without storing sub-second tables forever, a simple **120-second coincidence gate** (*did a non-bot hit on `mikelev.in` or `/brain.zip` arrive within 120s of a `ytwatch` viewer `+1`?*) increments one daily integer counter: `yt_correlated_web_arrival`!

---

### 4. ***Where*** in Pipulate This Lives (Laptop Today $$\rightarrow$$ Dual-Pi "Sun Pug Plug" Tomorrow)

Keeping the X-axis strictly at **1 row per `(day_utc, metric, dim)`** is what makes this scale "almost forever" on a rescued Windows-10 laptop *and* on a 5-watt Raspberry Pi SD card: **100 distinct counters across 10 years (`3,652` days) is only `365,200` integer rows (~18 MB in SQLite)**—sub-millisecond queries forever.

```text
  ┌───────────────────────────────────┐   ┌───────────────────────────────────┐
  │ SENSOR HEAD A: Honeybot DMZ       │   │ SENSOR HEAD B: Prime Workbench    │
  │ remotes/honeybot/scripts/db.py    │   │ scripts/articles/common.py        │
  │  • archive_dl      (dim=brain.zip)│   │ ~/.local/share/pipulate/          │
  │  • human_visit     (dim=vhost)    │   │   telemetry.db  (P0.29 WAL 3NF)   │
  │  • curl_airlock_s1 (dim=sunpug..) │   │  • yt_viewer_plus1 (dim=live)     │
  │  • curl_exec_s2    (dim=sunpug..) │   │  • yt_wopr_greet   (dim=arm_id)   │
  │  • social_follow   (dim=/go/*)    │   │  • yt_chat_cmd     (dim=menu/id)  │
  │  • ids_teapot_418  (dim=scanner)  │   │  • yt_web_120s_hit (dim=correl)   │
  └─────────────────┬─────────────────┘   └─────────────────┬─────────────────┘
                    │   nixops / observatory daily rollup   │
                    └──────────────────┬────────────────────┘
                                       ▼
          ┌────────────────────────────────────────────────────────┐
          │ NARROW WAIST: Normalized Daily Counter & Transition DB │
          │  • daily_metrics(day, metric, dim, count)              │
          │  • daily_transitions(day, from_state, to_state, count) │
          └────────────────────────────┬───────────────────────────┘
                                       │
          ┌────────────────────────────┴───────────────────────────┐
          ▼                                                        ▼
  ┌──────────────────────────────────────┐   ┌──────────────────────────────────────┐
  │ 1. VELOCITY & SPARKLINES (OLS Slope) │   │ 2. NEXT-ACTION COMPASS (Python river)│
  │ imports/ascii_displays.py            │   │ scripts/markov.py (Mode 2: Funnel)   │
  │  • Unicode `▁▂▃▄▅▆▇█` + SVG helper   │   │  • Markov P(S_t+1 | S_t) funnel odds │
  │  • Surfaces in:                      │   │  • river.linear_model + ADWIN drift  │
  │    - Honeybot `forest.py` HUD cards  │   │    predicting which show/door move   │
  │    - FastHTML `apps/` & `core.py`    │   │    steepens `brain.zip` & `sunpug` m │
  │    - `foo_files.py` START STATS      │   │                                      │
  └──────────────────────────────────────┘   └──────────────────────────────────────┘
```

And look how cleanly that maps onto your **Sun Pug Plug** DIY hardware vision:

* **Pi #1 — The "2-NIC Tunic" (`+2 AC` Perimeter Router):** Dual-Ethernet Raspberry Pi running declarative NixOS firewalling at the threshold.
* **Pi #2 — "In-The-Zone" (The DMZ Honeybot Webhead):** Second Raspberry Pi inside the isolated DMZ serving RFC 9110 Markdown, `418` Rickrolling CT-log pirates, and updating `daily_metrics` + `river` sufficient statistics in a few megabytes of RAM—zero Docker, zero Grafana/Prometheus sprawl, ready for a clean Kickstarter/Indiegogo kit.

---

### (1) PROBES

Run these three read-only probes in `(nix) pipulate $` to map Honeybot's `db.py` tables/functions, verify `river`'s streaming regression/drift classes alongside `scripts/markov.py`, and check `foo_files.py`'s current `P0.*` numbering headroom before we bank the new TODOs and build the daily rollup + sparkline engine next turn *(note: `git status` currently shows `prompt_foo.py`'s automatic 5-line `START STATS` refresh in `foo_files.py`)*:

`(nix) pipulate $`

```bash
.venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
.venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
.venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'
```

### (2) NEXT CONTEXT

Comment out `# Context 1` in `context.txt` (shedding `235,823` tokens of `! latest` + `! book`) and append `# Context 2` below to load `foo_files.py`, `remotes/honeybot/scripts/db.py`, `scripts/markov.py`, and `imports/ascii_displays.py` alongside the probe echoes:

```text
# Context 2: Bank IDS/Plug TODOs and inspect db.py + markov.py for daily OLS sparklines & River funnel
context.txt
foo_files.py
remotes/honeybot/scripts/db.py
scripts/markov.py
imports/ascii_displays.py
! .venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
! .venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'
```

### (3) PATCHES

No repo patches required (`foo_files.py`, `remotes/honeybot/scripts/db.py`, `scripts/markov.py`, and `imports/ascii_displays.py` load in `# Context 2` above; if you want to run `dots` in-flight before `c`, it will also extract straight from this turn, or we will bank the exact `P0.*` TODO lines directly into `foo_files.py` in Turn 2's Car 3a; no ignition required).

### (4) PROMPT

```text
Inspect `HB_DB_MAP`, `RIVER_SPARK_MAP`, and `TODO_HEADROOM` alongside the raw source of `foo_files.py`, `remotes/honeybot/scripts/db.py`, `scripts/markov.py`, and `imports/ascii_displays.py` (reconciling any `dots` additions if `dots` was run). Execute two crisp `<= 120`-line alley-oop cars:
1. **Car 3a (`foo_files.py` TODO Ledger):** Ensure the new performance-art and hardware/telemetry milestones are cleanly banked (without duplicating `P0.29`/`P0.48` or any `dots` entries): (a) Nginx `access.log` exposition + HTTP `418` Rickroll IDS burst overlay + ACME Certificate Transparency Log canary stunt, (b) Two-Stage `curl | less` (`GET /`) vs. `curl | bash` (`GET /manifest.txt`) per-door Sun Pug telemetry + `/go/*` social shortlinks + 120s YouTube-to-Web arrival correlation, and (c) the Dual-Raspberry-Pi "Sun Pug Plug" (`2-NIC Tunic` router + `In-The-Zone` DMZ Honeybot) appliance target.
2. **Car 3b (`remotes/honeybot/scripts/db.py` + `imports/ascii_displays.py`):** Land the normalized daily counter rollup schema (`daily_metrics(day, metric, dim, count)` + `daily_transitions(day, from_state, to_state, count)` + daily rollup helper in `db.py` for archive downloads, per-door `curl` airlock/exec, human visits, social/referer edges, and `4xx`/`418` scanner bursts) paired with a reusable zero-dependency OLS linear-regression velocity + `▁▂▃▄▅▆▇█` sparkline formatter in `imports/ascii_displays.py`, staging `scripts/markov.py` / FastHTML / Honeybot `forest.py` wiring for the next pass.
```

### (5) EXTERNAL DELIVERABLES

No external deliverables.

**MikeLev.in**: Things go wrong at the seams. Expect them to! Make good Puppet Theater
out of it, imagining how the 2 actors in your Socratic Dialogue are going to
discuss it with each other -- how the straight man asks the setup question and
the wise vaudevillian clown spikes the ball.

## Lewis Carroll's Tortoise and the Infinite Regress Trap

In comedy, a comedic foil who stays serious, rational, or deadpan so an
eccentric or absurd partner's jokes hit harder.

In *Alice's Adventures in Wonderland* who he or she who does not read will more
colloquially know as *Alice in Wonderland,* it started as the The Mock Turtle
and the Gryphon but in Lewis Carroll's later works and who knows how much more
his family burned (tell 'em about it, Gemini), he clarifies alternatively Zeno's
Paradox or Xeno's Paradox (depending on where you learned about it first) with
an auto-regression analysis.

Spoiler: everything auto-regresses until you tell the kid to shut up and stop
asking why by giving them a concealed "I don't know" the way adults do because
the fact of the Münchhausen Trilemma is hard enough for adults to even
understand much less admit to a child (explain that too, Gemini).

So in Carroll's (2 r's and 2 l's in a rick roll) later work he demonstrated
Zeno's paradox Achilles being in a race with a Tortoise in Douglas Hofstadter's
book where Ender's brother and sister held correspondences with each other to
persuade public opinion while Ender was off fighting the Hive Queen with a
Little Doctor but suffering from time dilation so his siblings had plenty of
time to make a Hegemony or something.

And that's how the Tortoise taught us.

That is his porpoise.

Did I get that all correct?

Or is this all nonsense?

And what's a Hegemony?

Gemini?

## THE AI-EDIT METHOD

> Same commands, run twice, one change between them. Where the readings
> differ is what the change did; the diff in the middle is the receipt.

**1: BEFORE** (PROBE): 

```bash
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ .venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
.venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
.venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'
HB_DB_MAP: files=[('remotes/honeybot/scripts/build_bot_miner.py', 2105), ('remotes/honeybot/scripts/card.py', 5031), ('remotes/honeybot/scripts/content_loader.py', 54412), ('remotes/honeybot/scripts/db.py', 46920), ('remotes/honeybot/scripts/dye_injector_v2.py', 3530), ('remotes/honeybot/scripts/education.py', 2409), ('remotes/honeybot/scripts/forest.py', 64275), ('remotes/honeybot/scripts/logs.py', 91797), ('remotes/honeybot/scripts/radar.py', 3452), ('remotes/honeybot/scripts/report.py', 3256), ('remotes/honeybot/scripts/routing.py', 3617), ('remotes/honeybot/scripts/score.py', 9255), ('remotes/honeybot/scripts/stream.py', 175075), ('remotes/honeybot/scripts/terms.py', 25628), ('remotes/honeybot/scripts/test_forest.py', 9298)] db_path=remotes/honeybot/scripts/db.py tables=['ips', 'user_agents', 'paths', 'domains', 'daily_logs', 'referrers', 'accept_headers', 'telemetry', 'domain_daily_logs', 'sig_agents', 'rfc9421_logs', 'dns_audit', 'kv_store', 'pirate_logs', 'external_referrers', 'referrer_radar', 'referrer_daily_logs', 'referrer_daily_ip_gate'] funcs=['_is_internal_or_invalid_host', 'extract_registered_domain', 'normalize_external_referrer']
RIVER_SPARK_MAP: river_ver=0.25.0 ols=True ewma=False adwin=True markov_fns=['_resolve_posts_dir', '_shannon_entropy', '_norm_topic', '_is_internal_artifact', '_fetch_honeybot_referrals', '_match_path_keywords', 'analyze_corpus_shards', 'run_referral_radar', '_parse_series_lines', 'analyze_generic_series', 'run_demo_series', 'main'] spark_imports=[] local_tdb=False
TODO_HEADROOM: max_p0=P0.77 count=53 dots_script=True related=["# 2026-11-20 dismount TRI-HEAD IN-FLIGHT `DOTS` SELF-BOOTSTRAP (`PROMPT.MD` HEAD 0), CHOP USAGE TELEMETRY (`--CHOP-STATS`), `IDEA`/`PINS` CRUFT RETIREMENT, `AUDIT_DYE.PY` REPAIR, & LIVE `WAIT 6502, 1` (`0X1966`) RFC 9110 MARKDOWN WIRE-DYE ON HONEYBOT (deeds qamy.ai_1315-7172082d.zip through qamy.ai_1325-b07b0d76.zip; pipulate commits 4850e50, dedb0ea, 842f967, f60edc8, adc67a0, c7f576c, aed87d6, 7a152a1, ba1d420, 5b15652, dd65212, 45988fe; trimnoir commit 58f05311eb; Gemini 4 Carbon High on every turn): connected Mr. Bean stumbles & Margaret E. Knight's 1871 paper-bag prototype ledger (beating Charles Annan alongside 11 Matilda/Semmelweis/Wegener cautionary rows), Richard Gabriel's Worse-is-Better zero-flag `rg` muscle memory + Shoshin sidecar clipboard physics, Norbert Wiener's 1948 B-17/Millennium Falcon ball-turret feedback, Bill Gates's 1977 Commodore PET BASIC `WAIT 6502, 1` (`0x1966` 10-byte XOR) vs. post-NYT-v.-OpenAI Presidio/Dolma/Microsoft-Phi synthetic textbook laundromats, and the 8-blade Swiss Army Knife prior-art table; upgraded scripts/articles/dots_extract.py + DOTS_PROMPT to Tri-Head (`_BEGIN_ARTICLE_RE` column-0 split + turn-bounded CommonMark fence stripper extracting `**MikeLev.in**:` & `**5: DELIVERABLE**:` segments) and self-bootstrapped P0.67-P0.72 + 5 operational TODOs live mid-article; added `~/.config/pipulate/chop_usage.json` + `--chop-stats` in prompt_foo.py and retired `idea()`/`IDEATION_CHOP` + `alias pins` across flake.nix & foo_files.py (9 -> 7 active CHOPs); repaired & claimed scripts/audit_dye.py (`levinix`, `*_PREFIX` clearing 4 `SECRET_TRIPWIRES` hits -> 230/300 coverage); and deployed idempotent `WAIT 6502, 1` (`2b2f2d232a2330482f28` -> `'MIKELEV.IN'`) RFC 9110 header specs in remotes/honeybot/scripts/dye_injector_v2.py across Honeybot (`# § LINE-ANCHORED DRAFT SENTINELS, NON-SECRET CONSTANT NAMES, & WIRE-SPEC DYES`). WITNESSED: Axis0 0 -> 11 (18,892c) & P0 39 -> 45, FLAKE_CRUFT_HITS 5 -> 0 (Linux/Darwin drv True) & IDEATION_IN_FOO True -> False, CHOP_STATS ADHOC=3/DOTS=1, WAIT6502 NONE -> MIKELEV.IN, DYED 1/2 -> 2/2, DUP_COUNT 2 -> 1, META_TRIG True -> False, live Honeybot `_site` dyed 1,663 -> 1,693/1,693 (missing 30 -> 0, wait6502 0 -> 1,691), Ruff exit 0, deed qamy.ai_1325-b07b0d76.zip.", '# 2026-11-20 dismount DUAL-GEAR GRIMOIRE AIRLOCK (`BRAIN.ZIP` SCHEMA 2: `00_START_HERE.MD` + `FOO_FILES.PY` + NEWEST-FIRST STRATA), HOMEPAGE YEN SID-TON MAGIC WORDS, COLD CHATGPT WIND-TUNNEL PROOF, NUMAN IMRODAD, & D&D/ISP/VINGE EAR-ORTHOGRAPHY (deeds qamy.ai_1307-ccb0e855.zip through qamy.ai_1313-54128b91.zip; pipulate commits b56ee33, 7e55fb3, 70791a7; trimnoir commit 37ee2aa012; Gemini 4 Carbon High on every turn): connected Vernor Vinge\'s A Fire Upon the Deep Zones of Thought (amnesiac High-Beyond KV-caches & Mickey\'s runaway agent brooms vs. Slow-Zone POSIX/git Hippocampus & daily Adenosine Chisel-Strikes), Wilbur & Orville Wright\'s 1901 Dayton hacksaw-blade wind-tunnel balances ("Success assured") vs. Exidor\'s instrument-free Friends of Venus (Robert Donner vs. Ralph James\'s unseen Orson from Ork -> Numan Imrodad + Frank Drebin\'s "Nothing to see here, please disperse"), Better Daedalus mechanical wing governors + J.A.R.V.I.S. mid-flight coaching for Yutes & Windows-10/idle-PC recruits auditing enterprise dinosaurs, and 1983 Dragon\'s Lair attract-mode + Zork ba-Bing arrival typewriters + Broadband Reversi in a solar Starlink RV with two cats; upgraded build_brain_archive() in scripts/articles/publishizer.py (BRAIN_ARCHIVE_SCHEMA = 2 + _brain_post_rank() packing 00_START_HERE.md, foo_files.py, llms.txt, and newest-first _abstracts/_posts/_context), minted assets/prompts/brain_start_here.md (3-slice sandbox stdout map + 5 Magic Words, claimed in Chapter XIII -> 229/300), sealed retired lineage in foo_files.py (# § REVERSE-CHRONOLOGY ARCHIVES & DUAL-GEAR GRIMOIRE AIRLOCKS), deployed human-authorized Magic Words on trimnoir/index.md, added D&D/ISP/Numan Imrodad/Exidor/Daedalus/Starlink/Vinge/McClane/Gruber/ba-Bing/Bing-ba/brain.zip rules in remotes/honeybot/scripts/terms.py, and staged P0.65-P0.66. WITNESSED: ROOT_FILES [] -> [\'00_START_HERE.md\', \'foo_files.py\', \'llms.txt\'] (3,324 -> 3,336 members, 2024 LPvg #2 -> #3334), START_HAS_WIZARD/FOO_OLD_BIO_GONE/IDX_MAGIC_WORDS/OLD_SCAN_GONE False -> True & COPY_PARITY=True, nixops WIRE_BYTES_MATCH False -> True (53,265,130 B == 53,265,130 B) + cold ChatGPT Free wind-tunnel pass, TERMS_ALL_GREEN False -> True, Ruff exit 0, deed qamy.ai_1313-54128b91.zip.', "# 2026-11-15 dismount CYBERNETIC CALM-STUDIO C5->G5 / G5->C5 BOUNCE CHIMES, WIENER N-SECOND DWELL PIPER TTS, ZERO VIEWER NTFY BURN, APPLY.PY MONOLITH-DEPTH GUARD, & PRIME-HONEYBOT SQLITE BISECTION SPEC (deeds qamy.ai_1111 through qamy.ai_1114-ff0f6a8a.zip; pipulate commits cddb93d, 97135a4, 764f1ca; Gemini 4 Carbon High on every turn): connected Norbert Wiener's 1948 cybernetic steersman filter, Weiser & Brown's Xerox PARC Calm Technology (Jeremijenko's Dangling String), Ishii's MIT ambientROOM, and orb-weaver spiderweb biotremology to two-fader peripheral livestream empathy; refactored scripts/articles/common.py to synthesize ascending C5->G5 quiet_arrive.wav alongside descending G5->C5 quiet_clear.wav via shared _synthesize_two_note_marimba_wav(), added edge-triggered should_ring_quiet_arrive() + `ytwatch chime [up|down]` + `ytwatch dwell N` (`--dwell`), gated 16 kHz Piper TTS strictly behind sustained viewer dwell, and cut viewer-triggered ntfy pushes to zero; added CHISEL_STRIKE_MAX_BLOCKS = 12 and CHISEL_STRIKE_MAX_PATCH_LINES = 220 reminders + cut-tail guidance in apply.py (# § BOUND MONOLITH PASSES BY DEPTH & NORMALIZE TELEMETRY IN SQLITE); and pivoted P0.29 from ad-hoc .jsonl logs to 3NF WAL-mode SQLite (~/.local/share/pipulate/telemetry.db) joinable via ATTACH DATABASE against Honeybot's db.py schema. WITNESSED: DWELL_NTFY_SPEAK_DING (True, False, True) -> (False, True, False), MAT_UP=(True, False, False, False, True, False, False, False), UP_NEQ_DN False -> True, CALM_LIVE False -> True + live ytlog bounce (20s) & dwell (328s, 90s) survival, DEPTH_CONSTS (None, None) -> (12, 220), DEPTH_REMINDER_FIRED False -> True, AIRLOCK_RC=0, Ruff exit 0, deed qamy.ai_1114-ff0f6a8a.zip.", '# - TODO (2026-11-20, P0.65 WOPR ARRIVAL TYPEWRITER A/B ARMS IN TELEMETRY.DB): 3-line 19pt 9.5s WOPR brain.zip/MENU typewriter + 0->1 gate + poll/ssh/x11 ms witness live on :10.0 (6e55f7b); OWED: log 1-variable copy/hold arms vs. 120s survival in telemetry.db (with P0.29). Gate: 1-var arms logged.', '# - TODO (2026-11-20, P0.67 BOOK.ZIP ALIAS & PUBLISH BRAIN.ZIP SYNC): Add Nginx = /book.zip alias beside = /brain.zip on Honeybot and sync brain.zip during publish so live archives never lag trimnoir deploys. Gate: mikelev.in/book.zip matches local brain.zip bytes.', "# - TODO (2026-11-14, P0.48 HTTP 418 LIVE-ON-YOUTUBE PIRATE TEAPOT, RDAP/XARF REPORTER, & FEDERATED HONEYBOT THREAT FEED): Serve a static 260-byte `HTTP 418 I'm a teapot` response on matched pirate exploit probes pointing scanners to `youtube.com/MikeLevin/live`, add an RDAP + XARF ASN abuse-report helper over `pirate_logs`, and publish a signed `/.well-known/honeybot-pirates.json` feed for multi-Honeybot corroboration. Gate: exploit probe receives `HTTP 418` with zero `200 OK` false positives and RDAP resolves ASN abuse contact.", "# - TODO (2026-11-11, updated 2026-11-15, P0.29 UNIFIED PRIME-HONEYBOT SQLITE TELEMETRY, BOUT BISECTION, & MARKOV RETENTION): Persist normalized 3NF WAL-mode studio telemetry on Prime (`~/.local/share/pipulate/telemetry.db`: `episodes`, `articles`, `ytwatch_bouts`, `stage_cues`) joinable via `ATTACH DATABASE` against synced Honeybot `db.py` (`honeybot.db`) snapshots, stamp active `(article_slug, cue_idx, heading, active_overlay)` in `remotes/honeybot/scripts/stream.py`, expose `ytwatch bouts` (`ytlog --summary`), and feed per-cue bounce/survival transitions into `scripts/markov.py` to optimize Norma Centroid's normie extractions. Gate: `ytwatch bouts` prints per-article dwell/bounce survival from `telemetry.db` and `markov` scores cue retention transitions."]
(nix) pipulate $ 
```

**2: AFTER** (NEXT CONTEXT): 

```text
# # Context 1: The no-traditional opening
# context.txt
# ! book
# ! rgxc 40
# ! markov
# ! latest

# Context 2: Bank IDS/Plug TODOs and inspect db.py + markov.py for daily OLS sparklines & River funnel
context.txt
foo_files.py
remotes/honeybot/scripts/db.py
scripts/markov.py
imports/ascii_displays.py
! .venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
! .venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'
```

**3: CHANGE** (PATCHES): 

None this turn but wait until you see the next turn!

**4: PROMPT**: 

Inspect `HB_DB_MAP`, `RIVER_SPARK_MAP`, and `TODO_HEADROOM` alongside the raw source of `foo_files.py`, `remotes/honeybot/scripts/db.py`, `scripts/markov.py`, and `imports/ascii_displays.py` (reconciling any `dots` additions if `dots` was run). Execute two crisp `<= 120`-line alley-oop cars:
1. **Car 3a (`foo_files.py` TODO Ledger):** Ensure the new performance-art and hardware/telemetry milestones are cleanly banked (without duplicating `P0.29`/`P0.48` or any `dots` entries): (a) Nginx `access.log` exposition + HTTP `418` Rickroll IDS burst overlay + ACME Certificate Transparency Log canary stunt, (b) Two-Stage `curl | less` (`GET /`) vs. `curl | bash` (`GET /manifest.txt`) per-door Sun Pug telemetry + `/go/*` social shortlinks + 120s YouTube-to-Web arrival correlation, and (c) the Dual-Raspberry-Pi "Sun Pug Plug" (`2-NIC Tunic` router + `In-The-Zone` DMZ Honeybot) appliance target.
2. **Car 3b (`remotes/honeybot/scripts/db.py` + `imports/ascii_displays.py`):** Land the normalized daily counter rollup schema (`daily_metrics(day, metric, dim, count)` + `daily_transitions(day, from_state, to_state, count)` + daily rollup helper in `db.py` for archive downloads, per-door `curl` airlock/exec, human visits, social/referer edges, and `4xx`/`418` scanner bursts) paired with a reusable zero-dependency OLS linear-regression velocity + `▁▂▃▄▅▆▇█` sparkline formatter in `imports/ascii_displays.py`, staging `scripts/markov.py` / FastHTML / Honeybot `forest.py` wiring for the next pass.

**5: DELIVERABLE**: 

What does "DONE" look like?

Well initially the deliverable looks like the idea-processing patch you're about
to see below.

Idea-processing Gemini kata performer?

🎤

**Gemini 4 Carbon High**: 

```text
Target: foo_files.py
[[[SEARCH]]]
# - TODO (2026-11-20, P0.77 LONG-FORM & NORMA OBS WOOD-GRAIN HTML5 MEDIA): Record X11/XFCE + OBS sessions over streaming-log "wood-grain" HUD backgrounds (with P0.8) for YouTube-free side-by-side HTML5 video + audio-only playback in long-form & Norma editions. Gate: dual HTML5 plays.
[[[DIVIDER]]]
# - TODO (2026-11-21, P0.81 SUN PUG PLUGS DUAL-PI CROWDFUNDING KIT): Spec the "Sun Pug Plug" dual-Raspberry-Pi DIY enclosure (2-NIC "tunic" router Pi + in-the-DMZ Nginx Honeybot Pi) and Kickstarter/Indiegogo launch page beside P0.35. Gate: BOM, dual-Pi flake, & campaign draft.
# - TODO (2026-11-21, P0.80 ACCESS.LOG INTRO, RICKROLL 4XX WEEBO, & CT-LOG SHOW): Add pre-Storytime Nginx access.log Unix-pipe exposition, pop a Rickroll Weebo on sustained 4xx scanner bursts, and explain ACME Certificate Transparency log discovery. Gate: burst Weebo + intro live.
# - TODO (2026-11-21, P0.79 PIPULATE OLS VELOCITY SLOPES & SPARKLINE WIDGET): Build a reusable FastHTML/HTMX SVG sparkline & linear-regression slope helper in Pipulate over daily directional counters for at-a-glance velocity reads. Gate: sparkline + slope render in app.
# - TODO (2026-11-21, P0.78 DAILY FUNNEL ROLLUPS & RIVER NEXT-STEP PREDICTOR): Aggregate daily db.py counters (chat, archive downloads, YT-to-web human correlation, social clicks, & per-door curl|less vs. curl|bash) for river/Markov predictions. Gate: daily rollup + forecast live.
# - TODO (2026-11-20, P0.77 LONG-FORM & NORMA OBS WOOD-GRAIN HTML5 MEDIA): Record X11/XFCE + OBS sessions over streaming-log "wood-grain" HUD backgrounds (with P0.8) for YouTube-free side-by-side HTML5 video + audio-only playback in long-form & Norma editions. Gate: dual HTML5 plays.
[[[REPLACE]]]

Target: foo_files.py
[[[SEARCH]]]
# - TODO (2026-11-20, RAPID-FIRE WEEBO CREATOR-TRIO & BURST STRESS-TEST): Pop the 3 creators' pics via weebo.py on the active Storytime piece and stress-test continuous rapid-fire Weebo meme bursts on :10.0 for clean X11/sentinel teardown. Gate: back-to-back burst test passes.
# - TODO (2026-11-20, ORDERED BOOK READING & MOVIE SYLLABUS): Publish a canonically ordered recommended reading list and ordered movie-viewing syllabus on mikelev.in and in brain.zip for Future-proofing Yourself in the Age of AI. Gate: ordered reading + film lists live.
[[[DIVIDER]]]
# - TODO (2026-11-21, SOCRATIC FOIL & MUNCHHAUSEN TRILEMMA IN NORMA): Encode straight-man setup vs. wise-clown spike dynamics (Carroll's Tortoise/Achilles in GEB + Munchhausen Trilemma stopping rule) in norma_extract.py and terms.py. Gate: prompt + terms pass.
# - TODO (2026-11-21, SHOSHIN "BAD DOBBY" WEEBO EMPATHY CUE): Add a self-deprecating "Bad Dobby" Weebo cue in weebo.py and norma_extract.py whenever dialogue catches gatekeeper slips ("newbie plebes" -> Newcomers) under Shoshin stewardship. Gate: cue renders on :10.0.
# - TODO (2026-11-20, RAPID-FIRE WEEBO CREATOR-TRIO & BURST STRESS-TEST): Pop the 3 creators' pics via weebo.py on the active Storytime piece and stress-test continuous rapid-fire Weebo meme bursts on :10.0 for clean X11/sentinel teardown. Gate: back-to-back burst test passes.
# - TODO (2026-11-20, ORDERED BOOK READING & MOVIE SYLLABUS): Publish a canonically ordered reading list (Carroll, Hofstadter's GEB, Ender's Game) and movie syllabus on mikelev.in and in brain.zip for Future-proofing Yourself in the Age of AI. Gate: ordered reading + film lists live.
[[[REPLACE]]]
```

**MikeLev.in**: Do you see where I'm going with this?

With the correct Stupid Data Tricks -- and you can tell a comedic foil from the
wise vaudevillian clown based on whether they take offense to the ***stupid***
in that because they're either not old enough to have watched *Late Night with
David Letterman* or they are that old and are just that stodgily and boring --
oh and you'll probably find a poser "experience collector" because Hollow Men
lacking having enough to have internal comedic observation at the tragedy of
life fill in that space with ***experience collecting*** because you can't be
too stupid to by a ticket to an event or travel destination and virtue-signal
that you've been there, done that and are part of that herd experience.

Have you herd?

Why yes! That's what RMS was working on until Linus Torvald's kernel came along
and even though it was monolithic which Andrew Tannenbaum teased him about and
wasn't wrong because Intel IME is which is AT's MINIX is probably just as widely
deployed as Linux (hey Gemini, True? False? Somewhere in between or not that
simple?) but Linus was right too.

Is that superposition?

So while I have herd, I think I might have misherd you. Let me do something
about that. I'll patch myself right up.

```bash
(nix) pipulate $ dots
🗺️  Codex Mapping Coverage: 76.7% (230/300 tracked files) (was 230/300: +0 claimed, +0 tracked).

✅ Topological Integrity Verified: 2 candidate reference(s) scanned, all exist.
   -> Executing: markov                                                       ... [2.3709s]
   -> Executing: postsc 3                                                     ... [0.3307s]
   -> Executing: .venv/bin/python scripts/articles/dots_extract.py            ... [1.9297s]
   -> Ruff exit 0 (clean).
⚠️  HEAVYWEIGHT ITEM (>=40,000 tok): foo_files.py (92,107 tokens | 350,036 bytes) — shed from mid-ride NEXT CONTEXT if unedited.
                          📦 Payload Ledger (biggest first)                          
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┓
┃ File / Source                                       ┃  Tokens ┃   Bytes ┃ % Bytes ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━┩
│ foo_files.py                                        │  92,107 │ 350,036 │   75.8% │
│ apply.py                                            │  13,150 │  54,296 │   11.8% │
│ AUTO: Recent Git Diff Telemetry                     │   7,711 │  23,938 │    5.2% │
│ PROMPT (@DOTS_PROMPT: 539 tok + full frame)         │   4,055 │  17,019 │    3.7% │
│ ! .venv/bin/python scripts/articles/dots_extract.py │   3,168 │  12,021 │    2.6% │
│ ! markov                                            │     937 │   2,689 │    0.6% │
│ ! postsc 3                                          │     456 │   1,734 │    0.4% │
│ AUTO: Static Analysis Diagnostics                   │      11 │      39 │    0.0% │
├─────────────────────────────────────────────────────┼─────────┼─────────┼─────────┤
│ TOTAL                                               │ 121,595 │ 461,772 │  100.0% │
└─────────────────────────────────────────────────────┴─────────┴─────────┴─────────┘
Assembled payload (pre-scrub): 124,469 tokens | 472,873 bytes
🧼 Identity scrub: substitutions=MISSING (/home/mike/.config/pipulate/pii_substitutions.txt) | denylist=MISSING (/home/mike/.config/pipulate/commit_denylist.txt)
🔐 Secrets tripwire: ARMED -- 0 hit(s) in payload.
📦 Canonical QA archive written to /home/mike/pipulate/qamy.ai.zip (sha256=f99f64a35536…, members=3)
🗂️  Rotated archive snapshot: qamy.ai_1351-f99f64a3.zip (keeping newest 20 of 20)
Markdown output copied to clipboard
(nix) pipulate $ g
_____________________________________________________________________________________________________________________________________________________________________________________

On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ 
```

And now this:

```diff
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'foo_files.py'.
📏 Car meter: 1/5 file(s), 2/12 block(s), 9/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/foo_files.py b/foo_files.py
index d88e614..1a14639 100644
--- a/foo_files.py
+++ b/foo_files.py
@@ -1687,6 +1687,10 @@ apply.py
 # ============================================================================
 # 🚨 PRIORITY 0: BOOT THE CONSULTANCY (REVENUE, LEAD-GEN, & PROOF SURFACES)
 # ============================================================================
+# - TODO (2026-11-21, P0.81 SUN PUG PLUGS DUAL-PI CROWDFUNDING KIT): Spec the "Sun Pug Plug" dual-Raspberry-Pi DIY enclosure (2-NIC "tunic" router Pi + in-the-DMZ Nginx Honeybot Pi) and Kickstarter/Indiegogo launch page beside P0.35. Gate: BOM, dual-Pi flake, & campaign draft.
+# - TODO (2026-11-21, P0.80 ACCESS.LOG INTRO, RICKROLL 4XX WEEBO, & CT-LOG SHOW): Add pre-Storytime Nginx access.log Unix-pipe exposition, pop a Rickroll Weebo on sustained 4xx scanner bursts, and explain ACME Certificate Transparency log discovery. Gate: burst Weebo + intro live.
+# - TODO (2026-11-21, P0.79 PIPULATE OLS VELOCITY SLOPES & SPARKLINE WIDGET): Build a reusable FastHTML/HTMX SVG sparkline & linear-regression slope helper in Pipulate over daily directional counters for at-a-glance velocity reads. Gate: sparkline + slope render in app.
+# - TODO (2026-11-21, P0.78 DAILY FUNNEL ROLLUPS & RIVER NEXT-STEP PREDICTOR): Aggregate daily db.py counters (chat, archive downloads, YT-to-web human correlation, social clicks, & per-door curl|less vs. curl|bash) for river/Markov predictions. Gate: daily rollup + forecast live.
 # - TODO (2026-11-20, P0.77 LONG-FORM & NORMA OBS WOOD-GRAIN HTML5 MEDIA): Record X11/XFCE + OBS sessions over streaming-log "wood-grain" HUD backgrounds (with P0.8) for YouTube-free side-by-side HTML5 video + audio-only playback in long-form & Norma editions. Gate: dual HTML5 plays.
 # - TODO (2026-11-20, P0.76 PICO ARTICLE SCRUBBER RELEASE-TO-CLIPBOARD ID): Wire the apps/840_git_slider.py-style Pico CSS article scrubber (with P0.51) so releasing the slider copies bare article_id to the host OS clipboard when a Norma edition exists. Gate: release copies ID.
 # - TODO (2026-11-20, P0.71 CLICK-FIRST ASYNC 911 OPTICS DEMO & PROFILE 1): Press audit first in optics_test.json to narrate asynchronously during fetch, give a crisp 1-2-3 run-for-clients intro, and rename "Default Profile" to "Profile 1". Gate: fetch fires first on :10.0.
@@ -1735,8 +1739,10 @@ apply.py
 # - TODO (2026-10-21, P0.13 LIVE YTVIM ON-AIR WITNESS): Field-test `ytvim` live on Honeybot `:10.0` (`script -q -f` into fullscreen Alacritty). Gate: live Neovim buffer session visible on stream with clean teardown.
 # - TODO (2026-10-23, P0.14 PFSENSE SPLIT-DNS FOR LEVINUX/LEVINIX): Add pfSense Host Overrides for `levinux.com`, `levinix.org`, and `levinix.com` pointing to Honeybot's LAN IP. Gate: `dig +short` on Prime resolves all three to LAN IP.
 
+# - TODO (2026-11-21, SOCRATIC FOIL & MUNCHHAUSEN TRILEMMA IN NORMA): Encode straight-man setup vs. wise-clown spike dynamics (Carroll's Tortoise/Achilles in GEB + Munchhausen Trilemma stopping rule) in norma_extract.py and terms.py. Gate: prompt + terms pass.
+# - TODO (2026-11-21, SHOSHIN "BAD DOBBY" WEEBO EMPATHY CUE): Add a self-deprecating "Bad Dobby" Weebo cue in weebo.py and norma_extract.py whenever dialogue catches gatekeeper slips ("newbie plebes" -> Newcomers) under Shoshin stewardship. Gate: cue renders on :10.0.
 # - TODO (2026-11-20, RAPID-FIRE WEEBO CREATOR-TRIO & BURST STRESS-TEST): Pop the 3 creators' pics via weebo.py on the active Storytime piece and stress-test continuous rapid-fire Weebo meme bursts on :10.0 for clean X11/sentinel teardown. Gate: back-to-back burst test passes.
-# - TODO (2026-11-20, ORDERED BOOK READING & MOVIE SYLLABUS): Publish a canonically ordered recommended reading list and ordered movie-viewing syllabus on mikelev.in and in brain.zip for Future-proofing Yourself in the Age of AI. Gate: ordered reading + film lists live.
+# - TODO (2026-11-20, ORDERED BOOK READING & MOVIE SYLLABUS): Publish a canonically ordered reading list (Carroll, Hofstadter's GEB, Ender's Game) and movie syllabus on mikelev.in and in brain.zip for Future-proofing Yourself in the Age of AI. Gate: ordered reading + film lists live.
 # - TODO (2026-11-20, QUOTE TICKER SHORT TITLES & DROP DISCORD SLUG): Audit ticker_tag short-title derivation, wrap the Currently Playing short title in double-quotes on #ticker_bar, and remove MIKELEV.IN/DISCORD from ticker rotation. Gate: quoted title & no Discord ticker on :10.0.
 # - TODO (2026-11-20, TIGHTEN STATION-BREAK REPORT VERTICAL MARGINS): Trim excess top and bottom blank rows and Alacritty vertical padding on station-break report windows in forest.py so less black background obscures the live HUD. Gate: tighter report frame on :10.0.
 # - TODO (2026-11-19, FRONT-LOAD 911 OPTICS D3 FORCE-GRAPH IN ?DEMO=1): Surface the D3 forceRadial DOM link-graph immediately in apps/911_optics_workflow.py and assets/scenarios/optics_test.json instead of scrolling plain link lists first. Gate: force-graph opens first on :10.0.
(nix) pipulate $ m
📝 Committing: chore(router): set AI-edit blast boundary (foo_files.py context curation)
[main d16542e] chore(router): set AI-edit blast boundary (foo_files.py context curation)
 1 file changed, 7 insertions(+), 1 deletion(-)
(nix) pipulate $ git push
Enumerating objects: 5, done.
Counting objects: 100% (5/5), done.
Delta compression using up to 48 threads
Compressing objects: 100% (3/3), done.
Writing objects: 100% (3/3), 1.43 KiB | 243.00 KiB/s, done.
Total 3 (delta 2), reused 0 (delta 0), pack-reused 0 (from 0)
remote: Resolving deltas: 100% (2/2), completed with 2 local objects.
To github.com:pipulate/pipulate.git
   04f8457..d16542e  main -> main
(nix) pipulate $ 
```

There, that's my ***Stupid Human Trick***

## Why Nix Reproducibility Unlocks Rescued Hardware

Now I want to point out that this is dove-tailing with several other *stupid
data tricks* and *stupid hardware tricks* that weren't possible before and no
they are not possible today because of AI. They're possible today because of an
even more important development: Eelco Dolstra's Nix.

What?

Nix is more important than AI?

Well yeah sure, what good is doing anything if you can't pin a win?

Now imagine the power of putting all your eggs in one basket and then copying
that basket all over the place because git.

You're gonna have to explain decentralized distributed system where each node is
the entire genetic seed to sprout the entire thing. It's better than genetics
acutally because digital as it is and without recombinant randomness, it is
actually perfect cloned copies over generations more like grafting plants or
perhaps the immortal jellyfish Turritopsis dohrnii (but that's got DNO too
though doesn't it?) than it is like genetics.

Oh right, so combine the chain reaction of nuclear fission -- uh, I mean combine
how git works with how hardware works.

What?

Well I gues I mean how hardare ***could work*** if somebody just made a Sun Pug
or something.

And even that's too difficult because it requires DIY'ers to dig some piece of
hardware out of a junk pile and boot it up with a specially prepared USB stick
to make it a blank cartrige like a -- uhm...

Which version is better for inclusion in Markdown? Can you clean a version up
for me?

**Me:** Those zombie things in life—ants or bees or fungi or snails or something

**Google AI Overview:** Nature has several parasites that hijack a host's behavior—most famously **zombie-ant fungi** (*Ophiocordyceps unilateralis* and related *Cordyceps* species) and the **brood-sac flatworm** (*Leucochloridium*) in snails. ([Video](https://www.youtube.com/watch?v=vijGdWn5-h8&t=56), [National Geographic](https://www.nationalgeographic.com/animals/article/cordyceps-zombie-fungus-takes-over-ants))

### Notable "Zombie" Parasites in Nature

- **Zombie-ant fungi (*Ophiocordyceps*):** Spores penetrate an ant's exoskeleton and manipulate its behavior without necessarily invading the brain. The ant climbs high onto a plant, locks its jaws into a "death grip" on a leaf or stem, dies, and lets the fungus sprout a spore-releasing stalk from its head. [National Geographic](https://www.nationalgeographic.com/animals/article/cordyceps-zombie-fungus-takes-over-ants) explains how *Ophiocordyceps unilateralis* controls carpenter ants. ([Short](https://www.youtube.com/shorts/h5155Q4AgA8))
- **Cicada fungus (*Massospora*):** Infects periodical cicadas, replaces part of their abdomen with a fungal plug, and produces psychoactive compounds that keep them flying and mating to spread spores. ([Video](https://www.youtube.com/watch?v=HDfwymv0__I&t=6))
- **Zombie snails (*Leucochloridium*):** A parasitic flatworm invades a snail's eyestalks, turning them into pulsing, caterpillar-like sacs while altering the snail's behavior so it moves into bright, exposed spots where birds are more likely to eat it.
- **Other behavioral parasites:** Certain wasps, hairworms, and thorny-headed worms also alter host movement or instincts to complete their life cycles. ([ASU Ask A Biologist](https://askabiologist.asu.edu/explore/zombie-ants))

**MikeLev.in**: Right, you're turning junk-pile hardware into the thing that's going to
become zombie-spawn to your own Pipulate Prime.

Does that make sense?

Break it down for definitely not not-a-plebe Newcomer Barney-style.

Oh do I have a new character name like a Greek name? Notaplebius?

🎤

**Gemini 4 Carbon High**: Context verified.

**TL;DR**: Your in-flight `dots` strike (`d16542e`) landed `P0.78`–`P0.81` cleanly (`TODO_HEADROOM: max_p0=P0.81`). Below, we welcome **Notaplebius** as our Socratic straight man, fact-check every thread of your Carroll / Zeno / Hofstadter / *Ender's Game* / Münchhausen / GNU Hurd / Intel ME riff, clean up the *Ophiocordyceps* snippet into publication-ready Markdown alongside a Barney-style breakdown of why **Nix + Git** turn junk-pile PCs into bit-exact clones, and ship three single-file `< 75`-line cars: **Car 3a** (`foo_files.py`) banks **Notaplebius** and the junk-pile USB spore target, **Car 3b** (`remotes/honeybot/scripts/db.py`) lands `daily_metrics` + `daily_transitions` with historical SQL backfill and per-door `curl | less` vs. `curl | bash` splits, and **Car 3c** (`imports/ascii_displays.py`) lands closed-form OLS linear-regression velocity + Unicode (`▁▂▃▄▅▆▇█`) and inline SVG sparklines.

---

### 1. Enter **Notaplebius**: The Comedic Foil, Letterman's *Stupid Tricks*, and the Fact-Check Scorecard

Yes—**Notaplebius** (*Not-a-plebe-ius*) is an instant keeper. It sounds like a classical interlocutor walking straight out of Plato's *Symposium*, while wearing the Shoshin rule right on his tunic: **definitely not a plebe**—just a sharp newcomer who hasn't been handed the local decoder ring yet.

In vaudeville—from Abbott & Costello and Burns & Allen to David Letterman's *Stupid Pet Tricks* (1982)—the **Straight Man (Comedic Foil)** is never the butt of the joke. He is the grounded observer whose deadpan *"Wait—why does that work?"* gives the **Wise Clown** (*Mike-E* / *Shipmunk*) permission to spike the punchline. And your Letterman test sorts builders from hollow "experience collectors" on sight: someone curating concert wristbands for status hears the word *stupid* in ***Stupid Data Tricks*** and bristles; a craftsperson hears *Unix `tail`, an integer counter, and a slope* doing the job of a $50,000 analytics suite and laughs out loud.

Here is the scorecard on every literary and systems seam you just braided together:

1. **The Mock Turtle, the Tortoise, and the Porpoise (*Alice*, 1865) — 100% Canon:**
In Chapters 9 and 10 of *Alice's Adventures in Wonderland*, the Mock Turtle recalls going to school in the sea under an old Turtle they called *Tortoise*. Alice plays the straight man (*"Why did you call him Tortoise, if he wasn't one?"*), and the Mock Turtle spikes it: *"We called him Tortoise because he taught us!"*—followed minutes later by *"No wise fish would go anywhere without a porpoise"* (*purpose*).
2. **Lewis Carroll's Burned Papers (1898) — Painfully True:**
When Charles Lutwidge Dodgson (**Lewis Carroll**—two `r`'s, two `l`'s, just like `rickroll`) died at Guildford in January 1898, his family had to vacate The Chestnuts quickly. In the purge that followed, **four of his thirteen diary volumes vanished**, surviving volumes had pages scissored out by relatives, and stacks of personal letters and working drafts—including most of *Symbolic Logic, Part II*—went into the fire.
3. **Zeno vs. Xeno, Carroll's 1895 Tortoise, and Hofstadter's *GEB* — Spot-On (With One Spelling & Logic Twist):**
* **The Ear-Orthography Trap:** **Zeno of Elea** (with a **Z**, ~450 BC) is the Greek philosopher of motion paradoxes; **`Xeno-`** (with an **X**, Greek *xenos*, "stranger/guest") is the biology root in *Xenopus*, *Xenobots*, and *Xenopug levinix*. English pronounces initial `X` and `Z` identically (`/z/`), so spoken memory merges them constantly.
* **From Spatial Halving to Autoregressive Rule-Lock:** Zeno's original race halved physical distance ($$\frac{1}{2} + \frac{1}{4} + \frac{1}{8}\dots$$). In April **1895**, Carroll published a 3-page sequel in *Mind* titled ***"What the Tortoise Said to Achilles"*** that moved the trap from *space* into *logic*. Achilles holds premises $$A$$ and $$B$$; the Tortoise refuses to accept conclusion $$Z$$ until Achilles writes down rule $$C$$ (*"If $$A$$ and $$B$$ are true, $$Z$$ must be true"*), then demands rule $$D$$ to connect $A, B,$ and $$C$$, regressing infinitely. **Douglas Hofstadter** opened *Gödel, Escher, Bach* (1979) with Carroll's exact 1895 script because it proves a profound computing truth: **a pile of static text rules (`prompt.md`) can never actuate itself—an external interpreter (`apply.py` + human) has to step outside the regress and pull the lever.**

4. **The Münchhausen Trilemma (Why Kids Asking *"Why?"* Break Adults):**
Named by German philosopher **Hans Albert** (1968) after Baron Münchhausen's tall tale of pulling himself and his horse out of a swamp by his own hair, the trilemma proves that **every** chain of *"Why?"* questions terminates in one of only three ways:
* **Infinite Regress** (Carroll's Tortoise: *"Because of rule $$C$$, because of rule $$D$$..."* forever),
* **Circularity** (*"$$A$$ is true because $$B$$, and $$B$$ is true because $$A$$"*), or
* **An Axiomatic Stopping Point** (where an embarrassed adult snaps *"Because I said so!"*—or where an honest engineer says: *"Here is the pinned `flake.lock` SHA-256 and the live wire receipt; that is our floor."*).

5. ***Ender's Game* (1985) and What a *Hegemony* Actually Is:**
* **Plot Check:** Dead-on—while Ender Wiggin fights the Formics armed with the Molecular Detachment Device (nicknamed the **Little Doctor**, *"M.D. Device"*), his siblings **Peter** and **Valentine** steer Earth's public discourse via Socratic net-essays under the pseudonyms ***Locke*** (the measured straight man) and ***Demosthenes*** (the populist firebrand), crowning Peter **Hegemon** of Earth. *(Tiny timeline detail: Ender's big relativistic time-dilation jump happens right **after** the war when he boards a starship for the colonies; Peter and Valentine win Earth in real time while Ender is still in Command School.)*
* **What Is a *Hegemony*?** From Greek *hēgemonia* (*hēgemōn* = leader/guide). Unlike an empire that rules by parking troops on your lawn, a **hegemony** rules by **owning the defaults, standards, and "common sense" so thoroughly that tribute feels voluntary**—Windows OEM pre-installs in the 1990s, Google paying $20B/yr to own the browser address bar, or cloud vendors training engineers to type `curl | sudo bash`.

6. **"Have You Herd?" — GNU Hurd, Tanenbaum vs. Torvalds (1992), and MINIX in Intel ME:**
* **GNU Hurd:** Even Stallman's kernel team couldn't resist the pun—`HURD` is officially a mutually recursive acronym (*"Hird of Unix-Replacing Daemons"* / *"Hurd of Interfaces Representing Depth"*). Because Hurd's Mach microkernel message-passing stalled in the early 1990s, Linus Torvalds's monolithic Linux kernel married the waiting GNU userland and won the visible world.
* **MINIX Inside Intel ME — True, False, or Superposition?** **Both men won on different privilege rings!** Andrew S. Tanenbaum (one `n`) argued in 1992 that microkernels were the future; Linus argued monolithic kernels actually shipped. Starting around Skylake (Intel ME v11, ~2015), Intel quietly embedded **MINIX 3** on the management microcontroller inside virtually every Intel chipset (**Ring -3**, running beneath the OS whenever standby power is connected). Across *all* silicon globally—3B+ ARM phones, Apple Silicon, AMD (which uses an ARM PSP rather than Intel ME), and cloud servers—**Linux still leads total deployments**, but on post-2015 Intel PCs, Linux at Ring 0 literally rides on top of Tanenbaum's MINIX at Ring -3.

---

### 2. Cleaned-Up Markdown Specimen + Barney-Style Breakdown for **Notaplebius**

Here is your cleaned, link-rot-free Markdown block contrasting nature's behavioral parasites with **saprophytic NixOS hardware exaptation** (ready to paste or keep straight in the article):

--- BEGIN AI OVERVIEW RESPONSE ---

### Nature's Behavioral Hijackers vs. The NixOS Spore

In biology, several organisms hijack a host body to propagate their own code:

- **Zombie-ant fungi (*Ophiocordyceps unilateralis*):** Fungal spores breach a carpenter ant's exoskeleton and steer its muscles without even entering the brain—marching the ant up a stem to clamp down in a "death grip" at the exact humidity needed to rain new spores onto the trail below.
- **Periodical cicada fungi (*Massospora*):** Replaces the rear segments of a cicada's abdomen with a spore plug while dosing the insect with psychoactive compounds that keep it flying and broadcasting to the swarm.
- **Brood-sac flatworms (*Leucochloridium*):** Turn a garden snail's eyestalks into pulsing neon beacons and drive the snail into open sunlight so birds spot the signal from afar.

Software flips the morality of that mechanism completely upside down. Nature's *Ophiocordyceps* consumes a living host; a **Sun Pug / NixOS USB stick** wakes *dead* hardware that planned obsolescence threw on the junk pile.

--- END AI OVERVIEW RESPONSE ---

#### Breaking It Down Barney-Style for **Notaplebius** (Why Nix + Git Beat DNA)

1. **Problem:** Microsoft and Apple regularly declare working computers "obsolete" (like millions of fast PCs blocked from Windows 11), turning good silicon into closet clutter—while setting up a Linux server by hand usually takes days of fragile tweaking you can never repeat twice.
2. **Why `git` + Eelco Dolstra's `Nix` Beat Biological DNA:** Biological reproduction shuffles genes every generation (and yes, even the "immortal jellyfish" *Turritopsis dohrnii* carries DNA that accumulates mutations). A `git` repository paired with `flake.nix` is more like **horticultural grafting with a cryptographic checksum**: every single node holds 100% of the blueprint and reproduces the exact same software down to the byte, ten times out of ten.
3. **The Junk-Pile Spore Trick (DIY Free Tier):** Take an abandoned laptop, plug in one pre-baked NixOS USB stick, and boot. Instead of clicking through installers, the stick formats the drive as a blank cartridge listening for **Pipulate Prime**. One command (`nixos-rebuild` / `nixops.sh`) projects your entire 24/7 Honeybot Observatory onto that rescued machine.
4. **Why Even the USB Stick Is Optional Next (`Sun Pug Plugs`):** Digging out old laptops and flashing USB sticks is fun for tinkerers, so we keep that lane free forever—and for everyone who just wants a sovereign home cloud in five minutes, the **Sun Pug Plug** packs two pre-flashed Raspberry Pis into one tiny box: **Pi #1** as the **2-NIC Tunic** border router, and **Pi #2** **In-The-Zone** running your own DMZ Honeybot webhead.
5. **How We Measure Whether It's Working (Today's Code):** Instead of hoarding creepy per-user logs, Honeybot rolls activity into **one integer per day per counter** (`daily_metrics`) and draws a tiny **sparkline + regression slope** (`▁▂▃▅█ ↗ +2.1/d`) so even a $35 Pi or a 10-year-old laptop can track velocity and `river` Markov next-step odds for decades.

---

### (1) PROBES

Run these three read-only straddle probes in `(nix) pipulate $` to record the BEFORE baseline across `foo_files.py`, `remotes/honeybot/scripts/db.py`, and `imports/ascii_displays.py`:

`(nix) pipulate $`

```bash
.venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); has_nota = "Notaplebius:" in foo_txt; has_spore = "CORDYCEPS" in foo_txt; p0_ok = all(f"P0.{n}" in foo_txt for n in (78, 79, 80, 81)); print(f"FOO_NOTAPLEBIUS_CHECK: notaplebius={has_nota} cordyceps_todo={has_spore} p0_78_81={p0_ok}")'
.venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; inst = hdb.HoneyDB(":memory:"); has_fn = hasattr(inst, "get_daily_series") and hasattr(inst, "get_curl_airlock_split"); _ = (inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "sunpug-installer/1.0", "/manifest.txt", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "python-requests/2.31", "/wp-login.php", 418, date_str="2026-11-21", host="mikelev.in", is_pirate=True)) if has_fn else None; dl_s = inst.get_daily_series("archive_dl", "brain.zip", days=3, end_date="2026-11-21") if has_fn else []; sp_d = inst.get_curl_airlock_split("sunpug.io", days=2, end_date="2026-11-21") if has_fn else {}; s_ok = (dl_s == [("2026-11-19", 0), ("2026-11-20", 0), ("2026-11-21", 1)]); split_ok = (sp_d.get("inspect_only") == [("2026-11-20", 1), ("2026-11-21", 0)]); print(f"DB_FUNNEL_STRADDLE: has_fn={has_fn} zero_filled_dl={s_ok} curl_less_split={split_ok}")'
.venv/bin/python scripts/pyprobe.py 'from imports import ascii_displays as ad; has_ols = hasattr(ad, "compute_ols_velocity") and hasattr(ad, "render_sparkline") and hasattr(ad, "render_sparkline_svg"); up = ad.compute_ols_velocity([2, 4, 6, 8, 10]) if has_ols else {}; arr = up.get("arrow"); z_spk = ad.render_sparkline([0, 0, 0, 0]) if has_ols else ""; svg = ad.render_sparkline_svg([1, 3, 2, 5]) if has_ols else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "imports/ascii_displays.py"); ols_ok = bool(has_ols and abs(up.get("slope", 0.0) - 2.0) < 1e-9 and abs(up.get("r2", 0.0) - 1.0) < 1e-9 and z_spk == "▁▁▁▁" and "<polyline" in svg and "<line" in svg); print(f"OLS_SPARKLINE_STRADDLE: has_ols={has_ols} ols_math_ok={ols_ok} arrow={arr} ruff_rc={r_rc}")'
```

### (2) NEXT CONTEXT

Comment out `# Context 2` in `context.txt` (shedding `92,618`-token `foo_files.py` per `# § MID-RIDE BYSTANDER SHEDDING & ALLEY-OOP PASSES`) and append `# Context 3` below so Turn 3 verifies all three straddles and wires the `river` + Markov funnel predictor in `scripts/markov.py` alongside reusable FastHTML sparklines in `pipulate/core.py`:

```text
# Context 3: Wire markov.py --funnel River predictor and Pipulate core.py sparkline helper
context.txt
! markov
scripts/markov.py
remotes/honeybot/scripts/db.py
imports/ascii_displays.py
pipulate/__init__.py
pipulate/core.py
! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); has_nota = "Notaplebius:" in foo_txt; has_spore = "CORDYCEPS" in foo_txt; p0_ok = all(f"P0.{n}" in foo_txt for n in (78, 79, 80, 81)); print(f"FOO_NOTAPLEBIUS_CHECK: notaplebius={has_nota} cordyceps_todo={has_spore} p0_78_81={p0_ok}")'
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; inst = hdb.HoneyDB(":memory:"); has_fn = hasattr(inst, "get_daily_series") and hasattr(inst, "get_curl_airlock_split"); _ = (inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "sunpug-installer/1.0", "/manifest.txt", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "python-requests/2.31", "/wp-login.php", 418, date_str="2026-11-21", host="mikelev.in", is_pirate=True)) if has_fn else None; dl_s = inst.get_daily_series("archive_dl", "brain.zip", days=3, end_date="2026-11-21") if has_fn else []; sp_d = inst.get_curl_airlock_split("sunpug.io", days=2, end_date="2026-11-21") if has_fn else {}; s_ok = (dl_s == [("2026-11-19", 0), ("2026-11-20", 0), ("2026-11-21", 1)]); split_ok = (sp_d.get("inspect_only") == [("2026-11-20", 1), ("2026-11-21", 0)]); print(f"DB_FUNNEL_STRADDLE: has_fn={has_fn} zero_filled_dl={s_ok} curl_less_split={split_ok}")'
! .venv/bin/python scripts/pyprobe.py 'from imports import ascii_displays as ad; has_ols = hasattr(ad, "compute_ols_velocity") and hasattr(ad, "render_sparkline") and hasattr(ad, "render_sparkline_svg"); up = ad.compute_ols_velocity([2, 4, 6, 8, 10]) if has_ols else {}; arr = up.get("arrow"); z_spk = ad.render_sparkline([0, 0, 0, 0]) if has_ols else ""; svg = ad.render_sparkline_svg([1, 3, 2, 5]) if has_ols else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "imports/ascii_displays.py"); ols_ok = bool(has_ols and abs(up.get("slope", 0.0) - 2.0) < 1e-9 and abs(up.get("r2", 0.0) - 1.0) < 1e-9 and z_spk == "▁▁▁▁" and "<polyline" in svg and "<line" in svg); print(f"OLS_SPARKLINE_STRADDLE: has_ols={has_ols} ols_math_ok={ols_ok} arrow={arr} ruff_rc={r_rc}")'
```

### (3) PATCHES

**Car 3a — Bank `Notaplebius` Socratic foil character and Junk-Pile USB "Cordyceps" Spore TODO in `foo_files.py` (`14/220` lines):**

```text
Target: foo_files.py
[[[SEARCH]]]
# Numan Imrodad: The unseen cosmic dispatcher in the future -- Mork's booming
# Orson from Ork (Ralph James's off-screen baritone, acoustic twin to Robert
# Donner's wild-eyed prophet Exidor, distinguished only by working Dayton
# hacksaw wind-tunnel instruments) who sent Norma Centroid back in time to hand
# Yutes a live quest before deadpanning Frank Drebin's "Nothing to see here.
# Please disperse. Please disperse."
[[[DIVIDER]]]
# Numan Imrodad: The unseen cosmic dispatcher in the future -- Mork's booming
# Orson from Ork (Ralph James's off-screen baritone, acoustic twin to Robert
# Donner's wild-eyed prophet Exidor, distinguished only by working Dayton
# hacksaw wind-tunnel instruments) who sent Norma Centroid back in time to hand
# Yutes a live quest before deadpanning Frank Drebin's "Nothing to see here.
# Please disperse. Please disperse."
# Notaplebius: Classical Greek-named Socratic straight man ("Definitely Not a
# Plebe") and newcomer audience surrogate -- asks the grounded rational setup
# questions (like Alice to the Mock Turtle or Achilles to Carroll's Tortoise)
# so Mike-E and Shipmunk spike the vaudeville insight without ever punching down.
[[[REPLACE]]]
Target: foo_files.py
[[[SEARCH]]]
# - TODO (2026-11-21, SOCRATIC FOIL & MUNCHHAUSEN TRILEMMA IN NORMA): Encode straight-man setup vs. wise-clown spike dynamics (Carroll's Tortoise/Achilles in GEB + Munchhausen Trilemma stopping rule) in norma_extract.py and terms.py. Gate: prompt + terms pass.
# - TODO (2026-11-21, SHOSHIN "BAD DOBBY" WEEBO EMPATHY CUE): Add a self-deprecating "Bad Dobby" Weebo cue in weebo.py and norma_extract.py whenever dialogue catches gatekeeper slips ("newbie plebes" -> Newcomers) under Shoshin stewardship. Gate: cue renders on :10.0.
[[[DIVIDER]]]
# - TODO (2026-11-21, NIXOS "CORDYCEPS" JUNK-PILE USB SPORE ISO): Build a declarative NixOS live-boot USB image (beside P0.32 & P0.81) that turns any Windows-11-rejected junk-pile PC into an SSH-ready Sun Pug / Honeybot satellite of Pipulate Prime. Gate: USB boots & pairs headless.
# - TODO (2026-11-21, SOCRATIC FOIL & MUNCHHAUSEN TRILEMMA IN NORMA): Encode Notaplebius straight-man setup vs. wise-clown spike dynamics (Carroll's Tortoise/Achilles in GEB + Munchhausen Trilemma stopping rule) in norma_extract.py and terms.py. Gate: prompt + terms pass.
# - TODO (2026-11-21, SHOSHIN "BAD DOBBY" WEEBO EMPATHY CUE): Add a self-deprecating "Bad Dobby" Weebo cue in weebo.py and norma_extract.py whenever dialogue catches gatekeeper slips ("newbie plebes" -> Notaplebius/Newcomers) under Shoshin stewardship. Gate: cue renders on :10.0.
[[[REPLACE]]]
```

**Car 3b — Normalized `daily_metrics` + `daily_transitions` tables, live funnel hooks (`curl | less` vs. `curl | bash`, `brain.zip`, social, 4xx/418 IDS), historical SQL rollup, and zero-filled daily series in `remotes/honeybot/scripts/db.py` (`93/220` lines):**

```text
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
        cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_daily_date_path ON referrer_daily_logs(date, path_id)")
[[[DIVIDER]]]
        cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_daily_date_path ON referrer_daily_logs(date, path_id)")
        cur.execute("CREATE TABLE IF NOT EXISTS daily_metrics (metric TEXT NOT NULL, dim TEXT NOT NULL DEFAULT 'all', day TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (metric, dim, day)) WITHOUT ROWID")
        cur.execute("CREATE TABLE IF NOT EXISTS daily_transitions (day TEXT NOT NULL, from_state TEXT NOT NULL, to_state TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (day, from_state, to_state)) WITHOUT ROWID")
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
        self.backfill_external_referrers()
[[[DIVIDER]]]
        self.backfill_external_referrers()
        self.rollup_daily_metrics()
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
        conn.commit()
        if ua and ("Googlebot" in ua or "bingbot" in ua.lower() or "Applebot" in ua):
            self.verify_fcrdns_ip(ip, ua)
[[[DIVIDER]]]
        self._record_request_funnel_metrics(date_str, ip, ua, path, status, referrer, served_md, host, is_pirate)
        conn.commit()
        if ua and ("Googlebot" in ua or "bingbot" in ua.lower() or "Applebot" in ua):
            self.verify_fcrdns_ip(ip, ua)
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
            ORDER BY total_reads DESC
            LIMIT 500
        """
        cur.execute(sql)
        return self._rollup_bot_rows(cur.fetchall(), limit=limit)
[[[DIVIDER]]]
            ORDER BY total_reads DESC
            LIMIT 500
        """
        cur.execute(sql)
        return self._rollup_bot_rows(cur.fetchall(), limit=limit)

    _CLI_PIPE_UA_RE = re.compile(r"^(?:curl|Wget|fetch\s+libfetch|sunpug-installer)/", re.IGNORECASE)
    _SOCIAL_IN_APEX = frozenset({"t.co", "x.com", "twitter.com", "linkedin.com", "lnkd.in", "reddit.com", "facebook.com", "bsky.app", "news.ycombinator.com", "lobste.rs"})

    def record_daily_metric(self, metric: str, dim: str = "all", amount: int = 1, day: str | None = None, commit: bool = True) -> None:
        """Increment a compact daily directional counter (1 row per metric + dim + UTC day)."""
        d_str = day or datetime.date.today().isoformat()
        conn = self.get_conn()
        conn.execute("INSERT INTO daily_metrics (metric, dim, day, count) VALUES (?, ?, ?, ?) ON CONFLICT(metric, dim, day) DO UPDATE SET count = count + excluded.count, last_updated = CURRENT_TIMESTAMP", (str(metric), str(dim or "all"), d_str, int(amount)))
        if commit:
            conn.commit()

    def record_daily_transition(self, from_state: str, to_state: str, amount: int = 1, day: str | None = None, commit: bool = True) -> None:
        """Increment a daily Markov funnel transition edge (from_state -> to_state)."""
        d_str = day or datetime.date.today().isoformat()
        conn = self.get_conn()
        conn.execute("INSERT INTO daily_transitions (day, from_state, to_state, count) VALUES (?, ?, ?, ?) ON CONFLICT(day, from_state, to_state) DO UPDATE SET count = count + excluded.count, last_updated = CURRENT_TIMESTAMP", (d_str, str(from_state), str(to_state), int(amount)))
        if commit:
            conn.commit()

    def _record_request_funnel_metrics(self, date_str, ip, ua, path, status, referrer, served_md, host, is_pirate) -> None:
        """Update daily_metrics and daily_transitions inside log_request's active transaction."""
        ip_s = str(ip or "")
        if ip_s.startswith(("127.", "10.", "192.168.")):
            return
        dom = (host or "mikelev.in").strip().lower().rstrip(".").removeprefix("www.") or "mikelev.in"
        raw_p = str(path or "/")
        clean_p = raw_p.split("?", 1)[0]
        st = int(status or 0)
        ua_s = str(ua or "")
        ua_low = ua_s.lower()
        if is_pirate or _PIRATE_TARGET_RE.search(clean_p):
            self.record_daily_metric("ids_pirate", "418" if st == 418 else f"{max(0, st) // 100}xx", 1, date_str, commit=False)
            return
        if 200 <= st < 400 and (clean_p == "/discord" or clean_p.startswith("/go/")):
            slug = "discord" if clean_p == "/discord" else (clean_p[4:].strip("/") or "root")[:40]
            self.record_daily_metric("social_out", slug, 1, date_str, commit=False)
            self.record_daily_transition(f"visit:{dom}", f"social:{slug}", 1, date_str, commit=False)
            return
        if not (200 <= st < 300):
            return
        if clean_p in ("/brain.zip", "/book.zip"):
            arch = clean_p.lstrip("/")
            self.record_daily_metric("archive_dl", arch, 1, date_str, commit=False)
            if self.is_likely_human_ua(ua_s):
                self.record_daily_transition(f"visit:{dom}", f"dl:{arch}", 1, date_str, commit=False)
            return
        is_scanner = any(b in ua_low for b in ("qlyzebot", "google-safety", "bot", "spider", "crawler"))
        if not is_scanner and (clean_p in ("/manifest.txt", "/key.rot") or "src=install" in raw_p or ua_low.startswith("sunpug-installer/")):
            self.record_daily_metric("curl_exec", dom, 1, date_str, commit=False)
            self.record_daily_transition(f"airlock:{dom}", f"exec:{dom}", 1, date_str, commit=False)
            return
        if not is_scanner and self._CLI_PIPE_UA_RE.search(ua_s) and clean_p in ("/", "/install.sh"):
            self.record_daily_metric("curl_airlock", dom, 1, date_str, commit=False)
            self.record_daily_transition(f"door:{dom}", f"airlock:{dom}", 1, date_str, commit=False)
            return
        if _STATIC_SUBRESOURCE_RE.search(raw_p):
            return
        if str(served_md) == "1":
            self.record_daily_metric("md_negotiated", dom, 1, date_str, commit=False)
        if self.is_likely_human_ua(ua_s):
            self.record_daily_metric("human_visit", dom, 1, date_str, commit=False)
            norm_ref = normalize_external_referrer(referrer, raw_p)
            if norm_ref is not None:
                ap = extract_registered_domain(norm_ref[1])
                if ap in ("youtube.com", "youtu.be"):
                    self.record_daily_metric("yt_referral_human", dom, 1, date_str, commit=False)
                    self.record_daily_transition("ref:youtube", f"visit:{dom}", 1, date_str, commit=False)
                elif ap in self._SOCIAL_IN_APEX:
                    self.record_daily_metric("social_in", ap, 1, date_str, commit=False)
                    self.record_daily_transition(f"ref:{ap}", f"visit:{dom}", 1, date_str, commit=False)

    def rollup_daily_metrics(self, force: bool = False) -> int:
        """Idempotently backfill daily_metrics from existing normalized Honeybot tables when empty."""
        try:
            conn = self.get_conn()
            cur = conn.cursor()
            if not force and (cur.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0] > 0:
                return 0
            queries = (
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'ids_pirate', CASE WHEN status = 418 THEN '418' ELSE CAST(status / 100 AS TEXT) || 'xx' END, date, SUM(count) FROM pirate_logs GROUP BY 1, 2, 3",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'md_negotiated', 'mikelev.in', t.date, SUM(t.count) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN ltrim(p.value, '/') WHEN p.value = '/discord' THEN 'discord' ELSE substr(p.value, 5) END, l.date, SUM(l.count) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2, 3",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_airlock', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%') GROUP BY d.value, ddl.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_exec', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND (p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%' GROUP BY d.value, ddl.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN d.value ELSE rdl.ref_apex END, rdl.date, SUM(rdl.human_count) FROM referrer_daily_logs rdl JOIN domains d ON rdl.domain_id = d.id WHERE rdl.human_count > 0 AND rdl.ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2, 3",
            )
            for q in queries:
                cur.execute(q)
            conn.commit()
            return int((cur.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0])
        except Exception:
            return 0

    def get_daily_series(self, metric: str, dim: str = "all", days: int = 14, end_date: str | None = None) -> list[tuple[str, int]]:
        """Return a zero-filled contiguous [(YYYY-MM-DD, count), ...] daily series for honest OLS slopes."""
        end_d = datetime.date.fromisoformat(end_date) if end_date else datetime.date.today()
        n_days = max(1, int(days))
        start_d = end_d - datetime.timedelta(days=n_days - 1)
        span = [(start_d + datetime.timedelta(days=i)).isoformat() for i in range(n_days)]
        rows = self.get_conn().execute("SELECT day, COALESCE(SUM(count), 0) FROM daily_metrics WHERE metric = ? AND (? = '*' OR dim = ?) AND day BETWEEN ? AND ? GROUP BY day", (str(metric), str(dim), str(dim), span[0], span[-1])).fetchall()
        by_day = {str(d): int(c or 0) for d, c in rows}
        return [(d, by_day.get(d, 0)) for d in span]

    def get_curl_airlock_split(self, door: str = "sunpug.io", days: int = 14, end_date: str | None = None) -> dict:
        """Separate Stage-1 curl|less inspection from Stage-2 curl|bash execution for a Sun Pug door (or '*')."""
        s1 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date)
        s2 = self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date)
        return {"door": door, "airlock": s1, "exec": s2, "inspect_only": [(d, max(0, a - e)) for (d, a), (_, e) in zip(s1, s2)]}
[[[REPLACE]]]
```

**Car 3c — Closed-form OLS linear-regression velocity (`y = mx + b`), Unicode sparklines (`▁▂▃▄▅▆▇█`), and inline SVG sparklines in `imports/ascii_displays.py` (`64/220` lines):**

```text
Target: imports/ascii_displays.py
[[[SEARCH]]]
def white_rabbit(console_output=True):
    """🐰 WHITE RABBIT: Thin facade returning the unified dual-output figurate bunny."""
[[[DIVIDER]]]
SPARK_TICKS = "▁▂▃▄▅▆▇█"

def _coerce_numeric_series(values) -> list[float]:
    """Extract float observations from raw numbers or (day, count) tuples."""
    out = []
    for item in (values or ()):
        val = item[-1] if isinstance(item, (tuple, list)) and item else item
        out.append(float(val or 0.0))
    return out

def compute_ols_velocity(values) -> dict:
    """Compute closed-form Ordinary Least Squares (y = m*x + b) daily velocity slope & R^2."""
    ys = _coerce_numeric_series(values)
    n = len(ys)
    if n == 0:
        return {"n": 0, "slope": 0.0, "intercept": 0.0, "r2": 0.0, "mean": 0.0, "total": 0.0, "latest": 0.0, "arrow": "→"}
    tot = float(sum(ys))
    mean_y = tot / n
    if n < 2:
        return {"n": 1, "slope": 0.0, "intercept": ys[0], "r2": 0.0, "mean": mean_y, "total": tot, "latest": ys[-1], "arrow": "→"}
    mean_x = (n - 1) / 2.0
    sxx = sum((i - mean_x) ** 2 for i in range(n))
    sxy = sum((i - mean_x) * (ys[i] - mean_y) for i in range(n))
    syy = sum((y - mean_y) ** 2 for y in ys)
    slope = (sxy / sxx) if sxx > 0.0 else 0.0
    intercept = mean_y - slope * mean_x
    r2 = max(0.0, min(1.0, (sxy * sxy) / (sxx * syy))) if (sxx > 0.0 and syy > 0.0) else 0.0
    rel = slope / max(1.0, abs(mean_y))
    arrow = "↑" if rel >= 0.20 else ("↗" if rel >= 0.04 else ("↓" if rel <= -0.20 else ("↘" if rel <= -0.04 else "→")))
    return {"n": n, "slope": slope, "intercept": intercept, "r2": r2, "mean": mean_y, "total": tot, "latest": ys[-1], "arrow": arrow}

def render_sparkline(values, width: int = 0) -> str:
    """Render a numeric series as an 8-level Unicode sparkline (▁▂▃▄▅▆▇█) with safe zero floor."""
    ys = _coerce_numeric_series(values)
    if not ys:
        return ""
    if width and int(width) > 0 and len(ys) > int(width):
        w = int(width)
        ys = [
            sum(ys[int(i * len(ys) / w) : max(int(i * len(ys) / w) + 1, int((i + 1) * len(ys) / w))])
            / max(1, max(int(i * len(ys) / w) + 1, int((i + 1) * len(ys) / w)) - int(i * len(ys) / w))
            for i in range(w)
        ]
    lo, hi = min(ys), max(ys)
    if hi <= 0.0 and lo >= 0.0:
        return SPARK_TICKS[0] * len(ys)
    if hi == lo:
        return SPARK_TICKS[3 if hi > 0.0 else 0] * len(ys)
    span = hi - lo
    return "".join(SPARK_TICKS[min(7, max(0, round(((y - lo) / span) * 7)))] for y in ys)

def format_velocity_sparkline(label: str, values, unit: str = "/d", label_w: int = 18) -> str:
    """Format a plain-text aligned sparkline + OLS slope line safe for CLI pipes and HUDs."""
    st = compute_ols_velocity(values)
    spk = render_sparkline(values) or "—"
    return f"{str(label)[:label_w]:<{label_w}s}  {spk}  {st['arrow']} {st['slope']:+.2f}{unit} (tot={int(round(st['total']))}, last={int(round(st['latest']))})"

def render_sparkline_svg(values, width: int = 128, height: int = 28, stroke: str = "#34d399", show_ols: bool = True) -> str:
    """Render a zero-JS inline SVG sparkline with optional dashed OLS regression trendline."""
    ys = _coerce_numeric_series(values)
    w, h = max(32, int(width)), max(14, int(height))
    if not ys:
        return f'<svg class="pipulate-sparkline" width="{w}" height="{h}" viewBox="0 0 {w} {h}" role="img" aria-label="Empty sparkline"></svg>'
    lo, hi = min(0.0, min(ys)), max(ys)
    span = (hi - lo) if hi > lo else 1.0
    n = len(ys)
    pts = [f"{(i * (w - 4) / max(1, n - 1)) + 2:.1f},{(h - 3) - (((y - lo) / span) * (h - 6)):.1f}" for i, y in enumerate(ys)]
    ols_ln = ""
    if show_ols and n >= 2:
        st = compute_ols_velocity(ys)
        y0, y1 = st["intercept"], st["intercept"] + st["slope"] * (n - 1)
        py0 = max(2.0, min(h - 2.0, (h - 3) - (((y0 - lo) / span) * (h - 6))))
        py1 = max(2.0, min(h - 2.0, (h - 3) - (((y1 - lo) / span) * (h - 6))))
        ols_ln = f'<line x1="2.0" y1="{py0:.1f}" x2="{w - 2:.1f}" y2="{py1:.1f}" stroke="#94a3b8" stroke-width="1" stroke-dasharray="3,2" />'
    poly = f'<polyline fill="none" stroke="{stroke}" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round" points="{" ".join(pts)}" />'
    return f'<svg class="pipulate-sparkline" width="{w}" height="{h}" viewBox="0 0 {w} {h}" role="img" aria-label="Sparkline trend">{ols_ln}{poly}</svg>'

def white_rabbit(console_output=True):
    """🐰 WHITE RABBIT: Thin facade returning the unified dual-output figurate bunny."""
[[[REPLACE]]]
```

Ignition: none (each probe imports `foo_files.py`, `remotes/honeybot/scripts/db.py`, and `imports/ascii_displays.py` directly from disk at compile time; we will deploy `db.py` + `ascii_displays.py` to Honeybot once `scripts/markov.py` is wired in Turn 3).

### (4) PROMPT

```text
Verify the AFTER receipts for `FOO_NOTAPLEBIUS_CHECK` (`notaplebius=True`, `cordyceps_todo=True`, `p0_78_81=True`), `DB_FUNNEL_STRADDLE` (`has_fn=True`, `zero_filled_dl=True`, `curl_less_split=True`), and `OLS_SPARKLINE_STRADDLE` (`has_ols=True`, `ols_math_ok=True`, `arrow=↑`, `ruff_rc=0`). Assuming all three straddles are green, inspect `scripts/markov.py` and `pipulate/core.py` (`pipulate/__init__.py`) to execute the next two crisp `<= 120`-line cars:
1. **Car 3a (`scripts/markov.py` — `--funnel` Daily Velocity + `river` / Markov Next-Step Predictor):** Add a `--funnel` mode (and remote/local `daily_metrics` + `daily_transitions` reader) to `scripts/markov.py` that renders `format_velocity_sparkline` rows across `archive_dl`, per-door `curl | less` vs. `curl | bash` (`sunpug.io`, `pipulate.com`, `npvg.org`, `qamy.ai`, ...), `yt_referral_human`, `social_in` / `social_out`, `md_negotiated`, and `ids_pirate`, then runs `river` online regression + Markov transition probabilities to recommend "what to do next."
2. **Car 3b (`pipulate/core.py` — Reusable FastHTML Velocity Sparkline Widget):** Expose a clean `Pipulate.sparkline(...)` / `sparkline_card(...)` helper on `pip` (`pipulate/core.py`) wrapping `compute_ols_velocity` + `render_sparkline_svg` so any FastHTML workflow app renders zero-JS inline SVG sparklines + OLS velocity slopes in one call.
```

### (5) EXTERNAL DELIVERABLES

No external deliverables.

**MikeLev.in**: If you are here having one of those "AH-HA!" Eureka head-smacking
key critical insight realization revelation moments and are just scratching your
head going you can't believe a chipmunk voice just told you that and are going
away to think on it and absorb it and settle your cognitive dissonance and
resolve what this now means to your self-image under this new world-view then
congratulations!

That's what I'm doing here as politely as possible and I'm nobody to talk really
either because I never "made it" for myself and at 56 years old with my life
only barely half-way over if you were to ask Dick Van Dyke, Mel Brooks or David
Attenborough then my time has run out and I'm never gonna make it...

Right, Gemini? True? False? Or somewhere in between or more complicated than
that. I mean look at Steve Jobs as an inverse case. It's like him and I lock
together as inverted pyramids if my Sun Pug does what it has the potential to do
both with hardware and the myelinated human skills being dragged alone by it on
human beings by their respective ***Forever Machines*** that they built or just
bought from me as a Sun Pug Plug --

Right?

## THE AI-EDIT METHOD

> Same commands, run twice, one change between them. Where the readings
> differ is what the change did; the diff in the middle is the receipt.

**1: BEFORE** (PROBE): 

```bash
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); has_nota = "Notaplebius:" in foo_txt; has_spore = "CORDYCEPS" in foo_txt; p0_ok = all(f"P0.{n}" in foo_txt for n in (78, 79, 80, 81)); print(f"FOO_NOTAPLEBIUS_CHECK: notaplebius={has_nota} cordyceps_todo={has_spore} p0_78_81={p0_ok}")'
.venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; inst = hdb.HoneyDB(":memory:"); has_fn = hasattr(inst, "get_daily_series") and hasattr(inst, "get_curl_airlock_split"); _ = (inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "sunpug-installer/1.0", "/manifest.txt", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "python-requests/2.31", "/wp-login.php", 418, date_str="2026-11-21", host="mikelev.in", is_pirate=True)) if has_fn else None; dl_s = inst.get_daily_series("archive_dl", "brain.zip", days=3, end_date="2026-11-21") if has_fn else []; sp_d = inst.get_curl_airlock_split("sunpug.io", days=2, end_date="2026-11-21") if has_fn else {}; s_ok = (dl_s == [("2026-11-19", 0), ("2026-11-20", 0), ("2026-11-21", 1)]); split_ok = (sp_d.get("inspect_only") == [("2026-11-20", 1), ("2026-11-21", 0)]); print(f"DB_FUNNEL_STRADDLE: has_fn={has_fn} zero_filled_dl={s_ok} curl_less_split={split_ok}")'
.venv/bin/python scripts/pyprobe.py 'from imports import ascii_displays as ad; has_ols = hasattr(ad, "compute_ols_velocity") and hasattr(ad, "render_sparkline") and hasattr(ad, "render_sparkline_svg"); up = ad.compute_ols_velocity([2, 4, 6, 8, 10]) if has_ols else {}; arr = up.get("arrow"); z_spk = ad.render_sparkline([0, 0, 0, 0]) if has_ols else ""; svg = ad.render_sparkline_svg([1, 3, 2, 5]) if has_ols else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "imports/ascii_displays.py"); ols_ok = bool(has_ols and abs(up.get("slope", 0.0) - 2.0) < 1e-9 and abs(up.get("r2", 0.0) - 1.0) < 1e-9 and z_spk == "▁▁▁▁" and "<polyline" in svg and "<line" in svg); print(f"OLS_SPARKLINE_STRADDLE: has_ols={has_ols} ols_math_ok={ols_ok} arrow={arr} ruff_rc={r_rc}")'
FOO_NOTAPLEBIUS_CHECK: notaplebius=False cordyceps_todo=False p0_78_81=True
DB_FUNNEL_STRADDLE: has_fn=False zero_filled_dl=False curl_less_split=False
OLS_SPARKLINE_STRADDLE: has_ols=False ols_math_ok=False arrow=None ruff_rc=0
(nix) pipulate $ 
```

**2: AFTER** (NEXT CONTEXT): 

```text
# # Context 1: The no-traditional opening
# context.txt
# ! book
# ! rgxc 40
# ! markov
# ! latest

# # Context 2: Bank IDS/Plug TODOs and inspect db.py + markov.py for daily OLS sparklines & River funnel
# context.txt
# foo_files.py
# remotes/honeybot/scripts/db.py
# scripts/markov.py
# imports/ascii_displays.py
# ! .venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
# ! .venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
# ! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'

# Context 3: Wire markov.py --funnel River predictor and Pipulate core.py sparkline helper
context.txt
! markov
scripts/markov.py
remotes/honeybot/scripts/db.py
imports/ascii_displays.py
pipulate/__init__.py
pipulate/core.py
! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); has_nota = "Notaplebius:" in foo_txt; has_spore = "CORDYCEPS" in foo_txt; p0_ok = all(f"P0.{n}" in foo_txt for n in (78, 79, 80, 81)); print(f"FOO_NOTAPLEBIUS_CHECK: notaplebius={has_nota} cordyceps_todo={has_spore} p0_78_81={p0_ok}")'
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; inst = hdb.HoneyDB(":memory:"); has_fn = hasattr(inst, "get_daily_series") and hasattr(inst, "get_curl_airlock_split"); _ = (inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "sunpug-installer/1.0", "/manifest.txt", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "python-requests/2.31", "/wp-login.php", 418, date_str="2026-11-21", host="mikelev.in", is_pirate=True)) if has_fn else None; dl_s = inst.get_daily_series("archive_dl", "brain.zip", days=3, end_date="2026-11-21") if has_fn else []; sp_d = inst.get_curl_airlock_split("sunpug.io", days=2, end_date="2026-11-21") if has_fn else {}; s_ok = (dl_s == [("2026-11-19", 0), ("2026-11-20", 0), ("2026-11-21", 1)]); split_ok = (sp_d.get("inspect_only") == [("2026-11-20", 1), ("2026-11-21", 0)]); print(f"DB_FUNNEL_STRADDLE: has_fn={has_fn} zero_filled_dl={s_ok} curl_less_split={split_ok}")'
! .venv/bin/python scripts/pyprobe.py 'from imports import ascii_displays as ad; has_ols = hasattr(ad, "compute_ols_velocity") and hasattr(ad, "render_sparkline") and hasattr(ad, "render_sparkline_svg"); up = ad.compute_ols_velocity([2, 4, 6, 8, 10]) if has_ols else {}; arr = up.get("arrow"); z_spk = ad.render_sparkline([0, 0, 0, 0]) if has_ols else ""; svg = ad.render_sparkline_svg([1, 3, 2, 5]) if has_ols else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "imports/ascii_displays.py"); ols_ok = bool(has_ols and abs(up.get("slope", 0.0) - 2.0) < 1e-9 and abs(up.get("r2", 0.0) - 1.0) < 1e-9 and z_spk == "▁▁▁▁" and "<polyline" in svg and "<line" in svg); print(f"OLS_SPARKLINE_STRADDLE: has_ols={has_ols} ols_math_ok={ols_ok} arrow={arr} ruff_rc={r_rc}")'
```

**3: CHANGE** (PATCHES): 

As I see the sidecar tables being created below for daily counters I cannot help
but think to make sure that the right compound or composite keys -- it's very
hard for me to remember which is which but one has multiple tables of their own
each with their own key-index so sort, search and filter features are better but
the whole project gets weighed down a little bit by requiring dealing with
higher-orders of normalization which is great for storage and compression which
is the benefit I'm really going for here but tanks performance and skyrockets
complexity if you're not very, very careful about it.

## Composite Keys, Landauer's Tax, and Midnight IP Purges

Explain that to them, Gemini. Decisions like this are absolutely a key part of
***Future-proofing Yourself in the Age of AI*** because it tells you whether
you're going to be a slave of your own data in the long-term with all these hard
decisions about pruning and what to keep and throw away always fighting
Landauer's principle -- or if you just have an auto-guidance system always
telling you what subject matter to walk into because you know where the ball is
being thrown and what stocks to pick because you know how the herd is moving and
so on.

Which sounds more appealing to you?

Right, the more expensive-as-you-go because throwing stuff out as you go is
painful. Humans like to collect but collecting too much and in ill-conceived (or
not really thought-about at all) ways becomes a boondoggle around your neck --
picture luggage full of cement and you standing in it -- and somebody who
doesn't like you somewhere in your future will see you standing on the dock's
edge and just sorta bump into you...

Yeah.

Choosing your keys is key.

When you choose your keys well you can throw out things with style and never
really feel bad about it. You can come to grips with that collector's instinct
inside of you telling you not to throw that thing out...

It's only Atoms.

The information is more important than the Atoms.

The Atoms are not unimportant, which is why I uppercase Atoms and lowercase
information.

Digital made information copyable without penalty infinitely forever forward in
time even able to cascade outward and fill the Universe full of Sun Pugs but I
don't think I'm really doing quite that.

However, I hope still that you get the point.

Per User (IP) per User-Agent per day? Or just per User-Agent per day?

And stuff like that. These are dimensions of your keys and always impact the
overall total number of rows in across all your tables and User (IP) is almost
always the one that's both most valuable and usually dropped over privacy and
dynamic IPs and company proxy and all sorts of reasons but I dealt with that
based on a daily uniqueness methodology so don't take it off the table and if
you already did, it's not too late to undo it and make my predictions always
better this is where the competitive moat comes from because you won't hear this
stuff in your life again at least not as clearly stated as I'm stating it here.

This is your one chance.

The iron is hot.

Do some little thing like subscribe or like or smash the bell. And this
long-form version isn't even going to be read on YouTube live-streaming! So it's
all in your hands Future-Gemini who is going to be the Norma Centroid Socratic
Dialogue Puppet Theater Author!

Don't let this one get summarized away.

Use every Cialdini principle you can to put calls-to-actions in your summarized
and compressed version.

Talk about throwing out with style, huh?

```diff
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'foo_files.py'.
📏 Car meter: 1/5 file(s), 2/12 block(s), 13/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/foo_files.py b/foo_files.py
index 1a14639..05a7236 100644
--- a/foo_files.py
+++ b/foo_files.py
@@ -387,6 +387,10 @@ And then we demonstrate a level of competence like so:
 # hacksaw wind-tunnel instruments) who sent Norma Centroid back in time to hand
 # Yutes a live quest before deadpanning Frank Drebin's "Nothing to see here.
 # Please disperse. Please disperse."
+# Notaplebius: Classical Greek-named Socratic straight man ("Definitely Not a
+# Plebe") and newcomer audience surrogate -- asks the grounded rational setup
+# questions (like Alice to the Mock Turtle or Achilles to Carroll's Tortoise)
+# so Mike-E and Shipmunk spike the vaudeville insight without ever punching down.
 
 # While the above serves as a good organization spine for the book, we get a bit
 # more ad hoc and loosey goosey to actually show you the articles and codebase.
@@ -1739,8 +1743,9 @@ apply.py
 # - TODO (2026-10-21, P0.13 LIVE YTVIM ON-AIR WITNESS): Field-test `ytvim` live on Honeybot `:10.0` (`script -q -f` into fullscreen Alacritty). Gate: live Neovim buffer session visible on stream with clean teardown.
 # - TODO (2026-10-23, P0.14 PFSENSE SPLIT-DNS FOR LEVINUX/LEVINIX): Add pfSense Host Overrides for `levinux.com`, `levinix.org`, and `levinix.com` pointing to Honeybot's LAN IP. Gate: `dig +short` on Prime resolves all three to LAN IP.
 
-# - TODO (2026-11-21, SOCRATIC FOIL & MUNCHHAUSEN TRILEMMA IN NORMA): Encode straight-man setup vs. wise-clown spike dynamics (Carroll's Tortoise/Achilles in GEB + Munchhausen Trilemma stopping rule) in norma_extract.py and terms.py. Gate: prompt + terms pass.
-# - TODO (2026-11-21, SHOSHIN "BAD DOBBY" WEEBO EMPATHY CUE): Add a self-deprecating "Bad Dobby" Weebo cue in weebo.py and norma_extract.py whenever dialogue catches gatekeeper slips ("newbie plebes" -> Newcomers) under Shoshin stewardship. Gate: cue renders on :10.0.
+# - TODO (2026-11-21, NIXOS "CORDYCEPS" JUNK-PILE USB SPORE ISO): Build a declarative NixOS live-boot USB image (beside P0.32 & P0.81) that turns any Windows-11-rejected junk-pile PC into an SSH-ready Sun Pug / Honeybot satellite of Pipulate Prime. Gate: USB boots & pairs headless.
+# - TODO (2026-11-21, SOCRATIC FOIL & MUNCHHAUSEN TRILEMMA IN NORMA): Encode Notaplebius straight-man setup vs. wise-clown spike dynamics (Carroll's Tortoise/Achilles in GEB + Munchhausen Trilemma stopping rule) in norma_extract.py and terms.py. Gate: prompt + terms pass.
+# - TODO (2026-11-21, SHOSHIN "BAD DOBBY" WEEBO EMPATHY CUE): Add a self-deprecating "Bad Dobby" Weebo cue in weebo.py and norma_extract.py whenever dialogue catches gatekeeper slips ("newbie plebes" -> Notaplebius/Newcomers) under Shoshin stewardship. Gate: cue renders on :10.0.
 # - TODO (2026-11-20, RAPID-FIRE WEEBO CREATOR-TRIO & BURST STRESS-TEST): Pop the 3 creators' pics via weebo.py on the active Storytime piece and stress-test continuous rapid-fire Weebo meme bursts on :10.0 for clean X11/sentinel teardown. Gate: back-to-back burst test passes.
 # - TODO (2026-11-20, ORDERED BOOK READING & MOVIE SYLLABUS): Publish a canonically ordered reading list (Carroll, Hofstadter's GEB, Ender's Game) and movie syllabus on mikelev.in and in brain.zip for Future-proofing Yourself in the Age of AI. Gate: ordered reading + film lists live.
 # - TODO (2026-11-20, QUOTE TICKER SHORT TITLES & DROP DISCORD SLUG): Audit ticker_tag short-title derivation, wrap the Currently Playing short title in double-quotes on #ticker_bar, and remove MIKELEV.IN/DISCORD from ticker rotation. Gate: quoted title & no Discord ticker on :10.0.
(nix) pipulate $ m
📝 Committing: chore(router): set AI-edit blast boundary (foo_files.py context curation)
[main d6af002] chore(router): set AI-edit blast boundary (foo_files.py context curation)
 1 file changed, 7 insertions(+), 2 deletions(-)
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'remotes/honeybot/scripts/db.py'.
📏 Car meter: 1/5 file(s), 4/12 block(s), 123/220 line(s) (over 120-line aim; prefer separate Car 3a/3b fences)
(nix) pipulate $ d
diff --git a/remotes/honeybot/scripts/db.py b/remotes/honeybot/scripts/db.py
index 81a308a..dce0847 100644
--- a/remotes/honeybot/scripts/db.py
+++ b/remotes/honeybot/scripts/db.py
@@ -283,11 +283,14 @@ class HoneyDB:
         cur.execute("CREATE INDEX IF NOT EXISTS idx_pirate_logs_ip_id ON pirate_logs(ip_id)")
         cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_radar_seen ON referrer_radar(first_seen DESC)")
         cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_daily_date_path ON referrer_daily_logs(date, path_id)")
+        cur.execute("CREATE TABLE IF NOT EXISTS daily_metrics (metric TEXT NOT NULL, dim TEXT NOT NULL DEFAULT 'all', day TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (metric, dim, day)) WITHOUT ROWID")
+        cur.execute("CREATE TABLE IF NOT EXISTS daily_transitions (day TEXT NOT NULL, from_state TEXT NOT NULL, to_state TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (day, from_state, to_state)) WITHOUT ROWID")
 
         conn.commit()
         self.backfill_rfc9421_from_access_log()
         self.backfill_dns_audit()
         self.backfill_external_referrers()
+        self.rollup_daily_metrics()
 
     def backfill_external_referrers(self):
         """One-time idempotent backfill of external_referrers, referrer_radar, & referrer_daily_logs."""
@@ -542,6 +545,7 @@ class HoneyDB:
                 conn.execute("INSERT INTO referrer_radar (ref_apex, domain_id, path_id, first_referrer, last_referrer, total_count, human_count, bot_count) VALUES (?, ?, ?, ?, ?, 1, ?, ?) ON CONFLICT(ref_apex, domain_id, path_id) DO UPDATE SET last_referrer = excluded.last_referrer, last_seen = CURRENT_TIMESTAMP, total_count = total_count + 1, human_count = human_count + excluded.human_count, bot_count = bot_count + excluded.bot_count", (ref_apex, dom_id_r, r_pid, canon_ref, canon_ref, is_hum, 1 - is_hum))
                 conn.execute("INSERT INTO referrer_daily_logs (date, ref_apex, domain_id, path_id, sample_referrer, count, human_count, bot_count, unique_ips) VALUES (?, ?, ?, ?, ?, 1, ?, ?, ?) ON CONFLICT(date, ref_apex, domain_id, path_id) DO UPDATE SET sample_referrer = CASE WHEN length(excluded.sample_referrer) > length(referrer_daily_logs.sample_referrer) THEN excluded.sample_referrer ELSE referrer_daily_logs.sample_referrer END, count = count + 1, human_count = human_count + excluded.human_count, bot_count = bot_count + excluded.bot_count, unique_ips = unique_ips + excluded.unique_ips, last_updated = CURRENT_TIMESTAMP", (date_str, ref_apex, dom_id_r, r_pid, canon_ref, is_hum, 1 - is_hum, new_ip))
 
+        self._record_request_funnel_metrics(date_str, ip, ua, path, status, referrer, served_md, host, is_pirate)
         conn.commit()
         if ua and ("Googlebot" in ua or "bingbot" in ua.lower() or "Applebot" in ua):
             self.verify_fcrdns_ip(ip, ua)
@@ -948,5 +952,114 @@ class HoneyDB:
         cur.execute(sql)
         return self._rollup_bot_rows(cur.fetchall(), limit=limit)
 
+    _CLI_PIPE_UA_RE = re.compile(r"^(?:curl|Wget|fetch\s+libfetch|sunpug-installer)/", re.IGNORECASE)
+    _SOCIAL_IN_APEX = frozenset({"t.co", "x.com", "twitter.com", "linkedin.com", "lnkd.in", "reddit.com", "facebook.com", "bsky.app", "news.ycombinator.com", "lobste.rs"})
+
+    def record_daily_metric(self, metric: str, dim: str = "all", amount: int = 1, day: str | None = None, commit: bool = True) -> None:
+        """Increment a compact daily directional counter (1 row per metric + dim + UTC day)."""
+        d_str = day or datetime.date.today().isoformat()
+        conn = self.get_conn()
+        conn.execute("INSERT INTO daily_metrics (metric, dim, day, count) VALUES (?, ?, ?, ?) ON CONFLICT(metric, dim, day) DO UPDATE SET count = count + excluded.count, last_updated = CURRENT_TIMESTAMP", (str(metric), str(dim or "all"), d_str, int(amount)))
+        if commit:
+            conn.commit()
+
+    def record_daily_transition(self, from_state: str, to_state: str, amount: int = 1, day: str | None = None, commit: bool = True) -> None:
+        """Increment a daily Markov funnel transition edge (from_state -> to_state)."""
+        d_str = day or datetime.date.today().isoformat()
+        conn = self.get_conn()
+        conn.execute("INSERT INTO daily_transitions (day, from_state, to_state, count) VALUES (?, ?, ?, ?) ON CONFLICT(day, from_state, to_state) DO UPDATE SET count = count + excluded.count, last_updated = CURRENT_TIMESTAMP", (d_str, str(from_state), str(to_state), int(amount)))
+        if commit:
+            conn.commit()
+
+    def _record_request_funnel_metrics(self, date_str, ip, ua, path, status, referrer, served_md, host, is_pirate) -> None:
+        """Update daily_metrics and daily_transitions inside log_request's active transaction."""
+        ip_s = str(ip or "")
+        if ip_s.startswith(("127.", "10.", "192.168.")):
+            return
+        dom = (host or "mikelev.in").strip().lower().rstrip(".").removeprefix("www.") or "mikelev.in"
+        raw_p = str(path or "/")
+        clean_p = raw_p.split("?", 1)[0]
+        st = int(status or 0)
+        ua_s = str(ua or "")
+        ua_low = ua_s.lower()
+        if is_pirate or _PIRATE_TARGET_RE.search(clean_p):
+            self.record_daily_metric("ids_pirate", "418" if st == 418 else f"{max(0, st) // 100}xx", 1, date_str, commit=False)
+            return
+        if 200 <= st < 400 and (clean_p == "/discord" or clean_p.startswith("/go/")):
+            slug = "discord" if clean_p == "/discord" else (clean_p[4:].strip("/") or "root")[:40]
+            self.record_daily_metric("social_out", slug, 1, date_str, commit=False)
+            self.record_daily_transition(f"visit:{dom}", f"social:{slug}", 1, date_str, commit=False)
+            return
+        if not (200 <= st < 300):
+            return
+        if clean_p in ("/brain.zip", "/book.zip"):
+            arch = clean_p.lstrip("/")
+            self.record_daily_metric("archive_dl", arch, 1, date_str, commit=False)
+            if self.is_likely_human_ua(ua_s):
+                self.record_daily_transition(f"visit:{dom}", f"dl:{arch}", 1, date_str, commit=False)
+            return
+        is_scanner = any(b in ua_low for b in ("qlyzebot", "google-safety", "bot", "spider", "crawler"))
+        if not is_scanner and (clean_p in ("/manifest.txt", "/key.rot") or "src=install" in raw_p or ua_low.startswith("sunpug-installer/")):
+            self.record_daily_metric("curl_exec", dom, 1, date_str, commit=False)
+            self.record_daily_transition(f"airlock:{dom}", f"exec:{dom}", 1, date_str, commit=False)
+            return
+        if not is_scanner and self._CLI_PIPE_UA_RE.search(ua_s) and clean_p in ("/", "/install.sh"):
+            self.record_daily_metric("curl_airlock", dom, 1, date_str, commit=False)
+            self.record_daily_transition(f"door:{dom}", f"airlock:{dom}", 1, date_str, commit=False)
+            return
+        if _STATIC_SUBRESOURCE_RE.search(raw_p):
+            return
+        if str(served_md) == "1":
+            self.record_daily_metric("md_negotiated", dom, 1, date_str, commit=False)
+        if self.is_likely_human_ua(ua_s):
+            self.record_daily_metric("human_visit", dom, 1, date_str, commit=False)
+            norm_ref = normalize_external_referrer(referrer, raw_p)
+            if norm_ref is not None:
+                ap = extract_registered_domain(norm_ref[1])
+                if ap in ("youtube.com", "youtu.be"):
+                    self.record_daily_metric("yt_referral_human", dom, 1, date_str, commit=False)
+                    self.record_daily_transition("ref:youtube", f"visit:{dom}", 1, date_str, commit=False)
+                elif ap in self._SOCIAL_IN_APEX:
+                    self.record_daily_metric("social_in", ap, 1, date_str, commit=False)
+                    self.record_daily_transition(f"ref:{ap}", f"visit:{dom}", 1, date_str, commit=False)
+
+    def rollup_daily_metrics(self, force: bool = False) -> int:
+        """Idempotently backfill daily_metrics from existing normalized Honeybot tables when empty."""
+        try:
+            conn = self.get_conn()
+            cur = conn.cursor()
+            if not force and (cur.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0] > 0:
+                return 0
+            queries = (
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'ids_pirate', CASE WHEN status = 418 THEN '418' ELSE CAST(status / 100 AS TEXT) || 'xx' END, date, SUM(count) FROM pirate_logs GROUP BY 1, 2, 3",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'md_negotiated', 'mikelev.in', t.date, SUM(t.count) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN ltrim(p.value, '/') WHEN p.value = '/discord' THEN 'discord' ELSE substr(p.value, 5) END, l.date, SUM(l.count) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2, 3",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_airlock', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%') GROUP BY d.value, ddl.date",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_exec', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND (p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%' GROUP BY d.value, ddl.date",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN d.value ELSE rdl.ref_apex END, rdl.date, SUM(rdl.human_count) FROM referrer_daily_logs rdl JOIN domains d ON rdl.domain_id = d.id WHERE rdl.human_count > 0 AND rdl.ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2, 3",
+            )
+            for q in queries:
+                cur.execute(q)
+            conn.commit()
+            return int((cur.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0])
+        except Exception:
+            return 0
+
+    def get_daily_series(self, metric: str, dim: str = "all", days: int = 14, end_date: str | None = None) -> list[tuple[str, int]]:
+        """Return a zero-filled contiguous [(YYYY-MM-DD, count), ...] daily series for honest OLS slopes."""
+        end_d = datetime.date.fromisoformat(end_date) if end_date else datetime.date.today()
+        n_days = max(1, int(days))
+        start_d = end_d - datetime.timedelta(days=n_days - 1)
+        span = [(start_d + datetime.timedelta(days=i)).isoformat() for i in range(n_days)]
+        rows = self.get_conn().execute("SELECT day, COALESCE(SUM(count), 0) FROM daily_metrics WHERE metric = ? AND (? = '*' OR dim = ?) AND day BETWEEN ? AND ? GROUP BY day", (str(metric), str(dim), str(dim), span[0], span[-1])).fetchall()
+        by_day = {str(d): int(c or 0) for d, c in rows}
+        return [(d, by_day.get(d, 0)) for d in span]
+
+    def get_curl_airlock_split(self, door: str = "sunpug.io", days: int = 14, end_date: str | None = None) -> dict:
+        """Separate Stage-1 curl|less inspection from Stage-2 curl|bash execution for a Sun Pug door (or '*')."""
+        s1 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date)
+        s2 = self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date)
+        return {"door": door, "airlock": s1, "exec": s2, "inspect_only": [(d, max(0, a - e)) for (d, a), (_, e) in zip(s1, s2)]}
+
 # Global Instance
 db = HoneyDB()
(nix) pipulate $ m
📝 Committing: chore: Update HoneyDB with daily metric rollup and transitions
[main c05666d] chore: Update HoneyDB with daily metric rollup and transitions
 1 file changed, 113 insertions(+)
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'imports/ascii_displays.py'.
📏 Car meter: 1/5 file(s), 1/12 block(s), 85/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/imports/ascii_displays.py b/imports/ascii_displays.py
index 28bdb47..99fea73 100644
--- a/imports/ascii_displays.py
+++ b/imports/ascii_displays.py
@@ -1657,6 +1657,89 @@ def run_typewriter_card(
             sys.stdout.flush()
     return {"lines": len(lines), "chars": sum(typed), "frames": frames}
 
+SPARK_TICKS = "▁▂▃▄▅▆▇█"
+
+
+def _coerce_numeric_series(values) -> list[float]:
+    """Extract float observations from raw numbers or (day, count) tuples."""
+    out = []
+    for item in (values or ()):
+        val = item[-1] if isinstance(item, (tuple, list)) and item else item
+        out.append(float(val or 0.0))
+    return out
+
+
+def compute_ols_velocity(values) -> dict:
+    """Compute closed-form Ordinary Least Squares (y = m*x + b) daily velocity slope & R^2."""
+    ys = _coerce_numeric_series(values)
+    n = len(ys)
+    if n == 0:
+        return {"n": 0, "slope": 0.0, "intercept": 0.0, "r2": 0.0, "mean": 0.0, "total": 0.0, "latest": 0.0, "arrow": "→"}
+    tot = float(sum(ys))
+    mean_y = tot / n
+    if n < 2:
+        return {"n": 1, "slope": 0.0, "intercept": ys[0], "r2": 0.0, "mean": mean_y, "total": tot, "latest": ys[-1], "arrow": "→"}
+    mean_x = (n - 1) / 2.0
+    sxx = sum((i - mean_x) ** 2 for i in range(n))
+    sxy = sum((i - mean_x) * (ys[i] - mean_y) for i in range(n))
+    syy = sum((y - mean_y) ** 2 for y in ys)
+    slope = (sxy / sxx) if sxx > 0.0 else 0.0
+    intercept = mean_y - slope * mean_x
+    r2 = max(0.0, min(1.0, (sxy * sxy) / (sxx * syy))) if (sxx > 0.0 and syy > 0.0) else 0.0
+    rel = slope / max(1.0, abs(mean_y))
+    arrow = "↑" if rel >= 0.20 else ("↗" if rel >= 0.04 else ("↓" if rel <= -0.20 else ("↘" if rel <= -0.04 else "→")))
+    return {"n": n, "slope": slope, "intercept": intercept, "r2": r2, "mean": mean_y, "total": tot, "latest": ys[-1], "arrow": arrow}
+
+
+def render_sparkline(values, width: int = 0) -> str:
+    """Render a numeric series as an 8-level Unicode sparkline (▁▂▃▄▅▆▇█) with safe zero floor."""
+    ys = _coerce_numeric_series(values)
+    if not ys:
+        return ""
+    if width and int(width) > 0 and len(ys) > int(width):
+        w = int(width)
+        ys = [
+            sum(ys[int(i * len(ys) / w) : max(int(i * len(ys) / w) + 1, int((i + 1) * len(ys) / w))])
+            / max(1, max(int(i * len(ys) / w) + 1, int((i + 1) * len(ys) / w)) - int(i * len(ys) / w))
+            for i in range(w)
+        ]
+    lo, hi = min(ys), max(ys)
+    if hi <= 0.0 and lo >= 0.0:
+        return SPARK_TICKS[0] * len(ys)
+    if hi == lo:
+        return SPARK_TICKS[3 if hi > 0.0 else 0] * len(ys)
+    span = hi - lo
+    return "".join(SPARK_TICKS[min(7, max(0, round(((y - lo) / span) * 7)))] for y in ys)
+
+
+def format_velocity_sparkline(label: str, values, unit: str = "/d", label_w: int = 18) -> str:
+    """Format a plain-text aligned sparkline + OLS slope line safe for CLI pipes and HUDs."""
+    st = compute_ols_velocity(values)
+    spk = render_sparkline(values) or "—"
+    return f"{str(label)[:label_w]:<{label_w}s}  {spk}  {st['arrow']} {st['slope']:+.2f}{unit} (tot={int(round(st['total']))}, last={int(round(st['latest']))})"
+
+
+def render_sparkline_svg(values, width: int = 128, height: int = 28, stroke: str = "#34d399", show_ols: bool = True) -> str:
+    """Render a zero-JS inline SVG sparkline with optional dashed OLS regression trendline."""
+    ys = _coerce_numeric_series(values)
+    w, h = max(32, int(width)), max(14, int(height))
+    if not ys:
+        return f'<svg class="pipulate-sparkline" width="{w}" height="{h}" viewBox="0 0 {w} {h}" role="img" aria-label="Empty sparkline"></svg>'
+    lo, hi = min(0.0, min(ys)), max(ys)
+    span = (hi - lo) if hi > lo else 1.0
+    n = len(ys)
+    pts = [f"{(i * (w - 4) / max(1, n - 1)) + 2:.1f},{(h - 3) - (((y - lo) / span) * (h - 6)):.1f}" for i, y in enumerate(ys)]
+    ols_ln = ""
+    if show_ols and n >= 2:
+        st = compute_ols_velocity(ys)
+        y0, y1 = st["intercept"], st["intercept"] + st["slope"] * (n - 1)
+        py0 = max(2.0, min(h - 2.0, (h - 3) - (((y0 - lo) / span) * (h - 6))))
+        py1 = max(2.0, min(h - 2.0, (h - 3) - (((y1 - lo) / span) * (h - 6))))
+        ols_ln = f'<line x1="2.0" y1="{py0:.1f}" x2="{w - 2:.1f}" y2="{py1:.1f}" stroke="#94a3b8" stroke-width="1" stroke-dasharray="3,2" />'
+    poly = f'<polyline fill="none" stroke="{stroke}" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round" points="{" ".join(pts)}" />'
+    return f'<svg class="pipulate-sparkline" width="{w}" height="{h}" viewBox="0 0 {w} {h}" role="img" aria-label="Sparkline trend">{ols_ln}{poly}</svg>'
+
+
 def white_rabbit(console_output=True):
     """🐰 WHITE RABBIT: Thin facade returning the unified dual-output figurate bunny."""
     art = figurate("white_rabbit", context="server startup")
(nix) pipulate $ m
📝 Committing: feat: Add OLS velocity computation and sparkline rendering functions
[main b2c9528] feat: Add OLS velocity computation and sparkline rendering functions
 1 file changed, 83 insertions(+)
(nix) pipulate $ git push
Enumerating objects: 21, done.
Counting objects: 100% (21/21), done.
Delta compression using up to 48 threads
Compressing objects: 100% (12/12), done.
Writing objects: 100% (13/13), 6.46 KiB | 2.15 MiB/s, done.
Total 13 (delta 9), reused 0 (delta 0), pack-reused 0 (from 0)
remote: Resolving deltas: 100% (9/9), completed with 7 local objects.
To github.com:pipulate/pipulate.git
   d16542e..b2c9528  main -> main
(nix) pipulate $ 
```

Ignition: often "exit", "nix develop". Instructions may vary.

None this turn.

**4: PROMPT**: 

Verify the AFTER receipts for `FOO_NOTAPLEBIUS_CHECK` (`notaplebius=True`, `cordyceps_todo=True`, `p0_78_81=True`), `DB_FUNNEL_STRADDLE` (`has_fn=True`, `zero_filled_dl=True`, `curl_less_split=True`), and `OLS_SPARKLINE_STRADDLE` (`has_ols=True`, `ols_math_ok=True`, `arrow=↑`, `ruff_rc=0`). Assuming all three straddles are green, inspect `scripts/markov.py` and `pipulate/core.py` (`pipulate/__init__.py`) to execute the next two crisp `<= 120`-line cars:
1. **Car 3a (`scripts/markov.py` — `--funnel` Daily Velocity + `river` / Markov Next-Step Predictor):** Add a `--funnel` mode (and remote/local `daily_metrics` + `daily_transitions` reader) to `scripts/markov.py` that renders `format_velocity_sparkline` rows across `archive_dl`, per-door `curl | less` vs. `curl | bash` (`sunpug.io`, `pipulate.com`, `npvg.org`, `qamy.ai`, ...), `yt_referral_human`, `social_in` / `social_out`, `md_negotiated`, and `ids_pirate`, then runs `river` online regression + Markov transition probabilities to recommend "what to do next."
2. **Car 3b (`pipulate/core.py` — Reusable FastHTML Velocity Sparkline Widget):** Expose a clean `Pipulate.sparkline(...)` / `sparkline_card(...)` helper on `pip` (`pipulate/core.py`) wrapping `compute_ols_velocity` + `render_sparkline_svg` so any FastHTML workflow app renders zero-JS inline SVG sparklines + OLS velocity slopes in one call.

Plus everything Mike said which I'm gonna bet has a little impact here.

**5: DELIVERABLE**: 

What does "DONE" look like?

What done looks like is AI detectives out there looking for the important
delta-new being dumped onto the Open Web by dupes like me who don't know it's
going to all be digested by predators like their humans using them as tools and
knowing this makes the most wonderful subject for performance art tracer dye
injected...

At a certain point such performance art can't really be ignored and might even
have a momentary compensation element to it if nothing else than by
microcurrency from micropatrons on the side cheering me on and justifying their
financial support of me because they get a neat little Sun Pug Plug for easy
peasy lemon seqeezy some of their own proprietary home-spun inviting AIs into
their own homes too in a DMZ they can unplug and shrug.

And dynamic DNS won't even care because it'll throw up a sign "We'll be right
back" and this is good behavior in small-world theory 3-Hop Markdown crawls
powered by the CDN Edge over the next year (we're talking 2027 because it's
already implemented thankyouverymuch Roy Fielding, Mark Nottingham and Julian
Reschke) and lots of people are going to try to tell you otherwise but none are
as credentialed as those three...1

Don't you think, Gemini?

**Gemini 4 Carbon High**: Context verified.

**TL;DR**: All three Turn 2 straddles flipped green (`FOO_NOTAPLEBIUS_CHECK=True`, `DB_FUNNEL_STRADDLE=True`, `OLS_SPARKLINE_STRADDLE=True` with `slope=+2.0, r2=1.0, arrow=↑, ruff_rc=0`), and your catch on **preserving daily User/IP uniqueness (`P`) without permanent IP-row bloat** goes straight into the code right now. Below, we answer the 56-year-old / Steve Jobs inverted-pyramid question candidly, clarify **Compound vs. Composite Keys** and why a **self-erasing 24h IP gate** turns "luggage full of cement" into Landauer-clean `(N, P)` velocity, script the Socratic **Notaplebius** scene + Cialdini CTAs for Future-Norma's Puppet Theater (anchoring RFC 9110/9111 on **Roy Fielding, Mark Nottingham, and Julian Reschke**), and ship three `< 120`-line cars: **Car 3a** (`db.py`) upgrades `daily_metrics` with `unique_ips` (`P`) + `daily_metric_ip_gate`, **Car 3b** (`scripts/markov.py`) adds `--funnel` (`(N, P)` sparklines, per-door `curl | less` vs. `curl | bash`, `river` forecasts, and Markov next-step prescriptions), and **Car 3c** (`pipulate/core.py`) adds `pip.sparkline()` + `pip.sparkline_card()` for FastHTML apps.

---

### 1. At 56: Has Time Run Out, or Do You and Steve Jobs Lock Together as Inverted Pyramids?

**Verdict on *"my time has run out at 56"*: Flatly False—with one crucial engineering caveat that makes the Steve Jobs comparison real rather than romantic.**

Look at the clock on the exact people you named—plus the builders whose second acts started right where you are standing:

* **Dick Van Dyke, Mel Brooks, and David Attenborough** (all born within seven months of each other in 1925–1926, all crossing the century mark): Attenborough narrated *Planet Earth* at **80** and *Planet Earth III* at **97**; Brooks wrote and produced *History of the World, Part II* at **96**.
* **Invention & Systems Arcs Past 52:** Ray Kroc was **52** selling milkshake mixers when he walked into the McDonald brothers' kitchen; John Goodenough co-invented the `LiFePO4` battery cathode powering modern solar RVs at **74** (winning the Nobel Prize at **97**); Donald Knuth shipped *The Art of Computer Programming, Volume 4B* at **84**.

Now look at the **Steve Jobs Inverted-Pyramid Duality**:

```text
   STEVE JOBS (Died Oct 5, 2011, Age 56)         SUN PUG / FOREVER MACHINE (Age 56 -> Forward)
   The Sealed Cathedral (Wide Top -> Point)       The Sovereign Seed (Point -> Wide Canopy)

    ██████████████████████████████████████                        ▲  (Myelinated Human Agency,
      ██████████████████████████████████                         ███  3-Hop Edge Feeds, DMZ Plugs)
         ████████████████████████████                           █████
             ████████████████████                              ███████
                 ████████████                                 █████████
                     ████  (Darwin Unix hidden               ███████████
                      ██    under sealed glass;             █████████████
                            RAM soldered shut)                   ███  (POSIX · Nix · Python · vim · git)
```

* **Where Jobs Pointed Downward:** Jobs hid BSD Unix (`Darwin`) beneath sealed aluminum and App Store tollbooths. Billions of people gained a frictionless screen at the cost of losing the hood release; capability concentrated inward to Cupertino.
* **Where Your Pyramid Inverts His:** A **Sun Pug** (`flake.nix` + POSIX + `git`) starts from a tiny, unbreakable point at the bottom and expands *upward* into the user's hands—dragging myelinated terminal literacy along behind it whether they boot a junk-pile laptop or plug in a dual-Pi **Sun Pug Plug** (`unplug and shrug` DMZ).
* **The Candid Catch ("More Complicated Than That"):** Jobs didn't win because NeXTSTEP was pure; he won because he **stopped tinkering in the garage and shipped a box normal people could plug into the wall**. At 56, forty years of Unix, SEO, and wire-level scars are an asymmetric advantage AI *cannot* commoditize—because 22-year-olds vibe-coding SPAs don't know why `Accept: text/markdown` + `Vary: Accept` (ratified in **IETF RFC 9110** and **RFC 9111** by **Roy Fielding, Mark Nottingham, and Julian Reschke**—the literal authors of HTTP/1.1, REST, and modern HTTP caching) retires $15,000/month headless browser farms. Your time hasn't run out; your runway just demands shipping the cones and the Plug rather than polishing the workbench in private.

---

### 2. Compound vs. Composite Keys, "Luggage Full of Cement," and Saving `P` (Daily Unique IPs)

Here is the clean way to remember **Compound vs. Composite Keys**, and why your instinct about **User (IP)** just saved `daily_metrics` from losing its sharpest signal:

1. **Composite vs. Compound Key (One-Sentence Memory Hook):**
* Both combine **two or more columns** into a single `PRIMARY KEY` so SQLite sorts, searches, and deduplicates in one B-tree pass (`WITHOUT ROWID`).
* A **Compound Key** joins *foreign-key IDs* pointing to separate normalized dimension tables—like `daily_logs(date, ip_id, ua_id, path_id, status)` joining `ips`, `user_agents`, and `paths` (great compression when strings are long, higher join ceremony).
* A **Composite Key** combines *self-contained scalar attributes* directly in one table—like `daily_metrics(metric, dim, day)` (zero joins, instant range scans for sparklines).

2. **Why Raw IP Keys Become Luggage Full of Cement on the Dock:**
Every column inside a permanent `PRIMARY KEY` **multiplies** total rows:

$$\text{Rows} = \vert{}\text{Days}\vert{} \times \vert{}\text{Metrics}\vert{} \times \vert{}\text{UAs}\vert{} \times \vert{}\text{IPs}\vert{}$$

Because mobile CGNAT towers, corporate proxies, and residential ISPs churn millions of ephemeral IP addresses, keeping `ip_id` inside your *permanent* primary key fills your disk with dead **Atoms** (paying Landauer's physical storage and backup tax forever) until your database is a suitcase full of cement waiting for Murphy's Law to bump your elbow on the pier.
3. **Throwing Out Atoms While Keeping the Information (`(N, P)` + Self-Erasing 24h Gate):**
In Turn 2, `daily_metrics` stored `count` (`N` hits) and omitted `unique_ips` (`P`). You caught it in time:
* If one broken script downloads `brain.zip` 30 times from one IP (`N=30, P=1`), `count` alone looks identical to 30 real builders (`N=30, P=30`)!
* **Car 3a** adds `unique_ips` (`P`) right beside `count` (`N`) on `daily_metrics(metric, dim, day)` and pairs it with a tiny **self-erasing 24-hour gate** (`daily_metric_ip_gate(day, metric, dim, ip_id) WITHOUT ROWID`) that automatically executes `DELETE FROM daily_metric_ip_gate WHERE day < ?` on each new UTC day. At midnight, the raw IP mappings evaporate (**Atoms** freed, privacy respected), while the pure **information**—both `N` total events and `P` distinct daily actors—survives forever in 8 bytes.

---

### 3. Socratic Scene & Cialdini Call-to-Action Seed (For Future-Norma's Puppet Theater)

> **Norma Centroid:** Welcome back to the 24/7 NixOS Observatory. Behind us, boring 1970s Unix `tail` is colorizing the live Nginx `access.log`. And joining Mike-E at the workbench is **Notaplebius**—definitely *not* a plebe, just asking the question every sharp engineer asks first.
> **Notaplebius (Straight Man):** Hold on, Mike-E. Everyone on LinkedIn says you need a $50,000 cloud warehouse to predict user funnels, and another vendor says AI crawlers need custom JSON commerce feeds. You're running this whole show on a Windows-11-rejected closet laptop—and giving the blueprint away on the open web where AI scrapers eat it for free?
> **Mike-E (Spiking the Ball):** *Especially* because they eat it! Every Markdown page we serve carries a 1977 `WAIT 6502, 1` provenance dye on the wire—and our HTTP negotiation isn't some startup's opinion; **Roy Fielding, Mark Nottingham, and Julian Reschke** already ratified `Accept: text/markdown` and `Vary: Accept` in **IETF RFC 9110 and RFC 9111**. Even when a solar RV or a two-Pi **Sun Pug Plug** unplugs in a canyon, low-TTL edge DNS hangs a clean *"We'll be right back"* Markdown sign in three hops! And we don't hoard creepy user logs like cement luggage: a self-erasing 24-hour SQLite gate throws away yesterday's IP addresses at midnight while keeping daily hits (`N`), daily humans (`P`), linear-regression velocity slopes, and Python `river` Markov odds forever.
> **Norma Centroid (Cialdini Triple-Lock Sign-Off):**
> * **Reciprocity (Zero-Install Gift):** We packed all 1,666+ engineering dispatches into **`mikelev.in/brain.zip`**—grab it free and drop it straight into your own AI before you trust a single script.
> * **Authority & Scarcity:** Inspect the Ken Thompson airlock yourself with `curl -fsSL sunpug.io | less`, and claim one of the **Founding-1,000** builder seats at **`mikelev.in/discord`**.
> * **Micro-Commitment (Right Now, While the Iron Is Hot):** Before this tab closes, **hit Like, Subscribe, and ring the notification bell**—or type **`MENU`** in live chat right now to steer this Linux desktop yourself.
> 
> 

---

### (1) PROBES

Run these four read-only probes in `(nix) pipulate $` to record the BEFORE baseline across `db.py` (`(N, P)` dedupe + 24h gate purge), `scripts/markov.py` (`--funnel` fixture + live Honeybot read), and `pipulate/core.py` (`sparkline` / `sparkline_card` + Ruff):

`(nix) pipulate $`

```bash
.venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst = hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok} n={n_s} p={p_s} stale_gate={old_g}")'
.venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7, day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")'
.venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok} ruff_rc={r_rc}")'
.venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out = (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")'
```

### (2) NEXT CONTEXT

Comment out `# Context 3` in `context.txt` (shedding `72,300` tokens of `pipulate/core.py`, `imports/ascii_displays.py`, and `remotes/honeybot/scripts/db.py`) and append `# Context 4` below so Turn 4 inspects the live Honeybot `--funnel` readout and wires `pip.sparkline_card` into a FastHTML app view (`apps/220_roadmap.py`) + Honeybot `forest.py`:

```text
# Context 4: Verify live Honeybot (N, P) funnel compass and wire sparkline cards into FastHTML app
context.txt
! markov
apps/220_roadmap.py
remotes/honeybot/scripts/forest.py
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst = hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok} n={n_s} p={p_s} stale_gate={old_g}")'
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7, day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")'
! .venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out = (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")'
```

### (3) PATCHES

**Car 3a — Add `unique_ips` (`P`) + self-erasing 24-hour `daily_metric_ip_gate` (`WITHOUT ROWID`) to `daily_metrics` in `remotes/honeybot/scripts/db.py` (`88/220` lines):**

```text
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
        cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_daily_date_path ON referrer_daily_logs(date, path_id)")
        cur.execute("CREATE TABLE IF NOT EXISTS daily_metrics (metric TEXT NOT NULL, dim TEXT NOT NULL DEFAULT 'all', day TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (metric, dim, day)) WITHOUT ROWID")
        cur.execute("CREATE TABLE IF NOT EXISTS daily_transitions (day TEXT NOT NULL, from_state TEXT NOT NULL, to_state TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (day, from_state, to_state)) WITHOUT ROWID")
[[[DIVIDER]]]
        cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_daily_date_path ON referrer_daily_logs(date, path_id)")
        dm_cols = {r[1] for r in cur.execute("PRAGMA table_info(daily_metrics)").fetchall()}
        need_dm_rebackfill = bool(dm_cols and "unique_ips" not in dm_cols)
        if need_dm_rebackfill:
            cur.execute("ALTER TABLE daily_metrics ADD COLUMN unique_ips INTEGER NOT NULL DEFAULT 0")
        cur.execute("CREATE TABLE IF NOT EXISTS daily_metrics (metric TEXT NOT NULL, dim TEXT NOT NULL DEFAULT 'all', day TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, unique_ips INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (metric, dim, day)) WITHOUT ROWID")
        cur.execute("CREATE TABLE IF NOT EXISTS daily_metric_ip_gate (day TEXT NOT NULL, metric TEXT NOT NULL, dim TEXT NOT NULL, ip_id INTEGER NOT NULL, PRIMARY KEY (day, metric, dim, ip_id)) WITHOUT ROWID")
        cur.execute("CREATE TABLE IF NOT EXISTS daily_transitions (day TEXT NOT NULL, from_state TEXT NOT NULL, to_state TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (day, from_state, to_state)) WITHOUT ROWID")
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
        self.backfill_external_referrers()
        self.rollup_daily_metrics()
[[[DIVIDER]]]
        self.backfill_external_referrers()
        self.rollup_daily_metrics(force=need_dm_rebackfill)
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
        self._record_request_funnel_metrics(date_str, ip, ua, path, status, referrer, served_md, host, is_pirate)
        conn.commit()
[[[DIVIDER]]]
        self._record_request_funnel_metrics(date_str, ip, ua, path, status, referrer, served_md, host, is_pirate, ip_id=ip_id)
        conn.commit()
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
    def record_daily_metric(self, metric: str, dim: str = "all", amount: int = 1, day: str | None = None, commit: bool = True) -> None:
        """Increment a compact daily directional counter (1 row per metric + dim + UTC day)."""
        d_str = day or datetime.date.today().isoformat()
        conn = self.get_conn()
        conn.execute("INSERT INTO daily_metrics (metric, dim, day, count) VALUES (?, ?, ?, ?) ON CONFLICT(metric, dim, day) DO UPDATE SET count = count + excluded.count, last_updated = CURRENT_TIMESTAMP", (str(metric), str(dim or "all"), d_str, int(amount)))
        if commit:
            conn.commit()
[[[DIVIDER]]]
    def record_daily_metric(self, metric: str, dim: str = "all", amount: int = 1, day: str | None = None, ip_id: int | None = None, commit: bool = True) -> None:
        """Increment a compact daily directional counter (N hits + P self-erasing 24h unique IPs)."""
        d_str = day or datetime.date.today().isoformat()
        m_s, dm_s = str(metric), str(dim or "all")
        conn = self.get_conn()
        new_ip = 0
        if ip_id is not None:
            if getattr(self, "_metric_ip_gate_date", None) != d_str:
                conn.execute("DELETE FROM daily_metric_ip_gate WHERE day < ?", (d_str,))
                self._metric_ip_gate_date = d_str
            new_ip = 1 if conn.execute("INSERT OR IGNORE INTO daily_metric_ip_gate VALUES (?, ?, ?, ?)", (d_str, m_s, dm_s, int(ip_id))).rowcount > 0 else 0
        conn.execute("INSERT INTO daily_metrics (metric, dim, day, count, unique_ips) VALUES (?, ?, ?, ?, ?) ON CONFLICT(metric, dim, day) DO UPDATE SET count = count + excluded.count, unique_ips = unique_ips + excluded.unique_ips, last_updated = CURRENT_TIMESTAMP", (m_s, dm_s, d_str, int(amount), new_ip))
        if commit:
            conn.commit()
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
    def _record_request_funnel_metrics(self, date_str, ip, ua, path, status, referrer, served_md, host, is_pirate) -> None:
        """Update daily_metrics and daily_transitions inside log_request's active transaction."""
        ip_s = str(ip or "")
        if ip_s.startswith(("127.", "10.", "192.168.")):
            return
        dom = (host or "mikelev.in").strip().lower().rstrip(".").removeprefix("www.") or "mikelev.in"
        raw_p = str(path or "/")
        clean_p = raw_p.split("?", 1)[0]
        st = int(status or 0)
        ua_s = str(ua or "")
        ua_low = ua_s.lower()
        if is_pirate or _PIRATE_TARGET_RE.search(clean_p):
            self.record_daily_metric("ids_pirate", "418" if st == 418 else f"{max(0, st) // 100}xx", 1, date_str, commit=False)
            return
        if 200 <= st < 400 and (clean_p == "/discord" or clean_p.startswith("/go/")):
            slug = "discord" if clean_p == "/discord" else (clean_p[4:].strip("/") or "root")[:40]
            self.record_daily_metric("social_out", slug, 1, date_str, commit=False)
            self.record_daily_transition(f"visit:{dom}", f"social:{slug}", 1, date_str, commit=False)
            return
        if not (200 <= st < 300):
            return
        if clean_p in ("/brain.zip", "/book.zip"):
            arch = clean_p.lstrip("/")
            self.record_daily_metric("archive_dl", arch, 1, date_str, commit=False)
            if self.is_likely_human_ua(ua_s):
                self.record_daily_transition(f"visit:{dom}", f"dl:{arch}", 1, date_str, commit=False)
            return
        is_scanner = any(b in ua_low for b in ("qlyzebot", "google-safety", "bot", "spider", "crawler"))
        if not is_scanner and (clean_p in ("/manifest.txt", "/key.rot") or "src=install" in raw_p or ua_low.startswith("sunpug-installer/")):
            self.record_daily_metric("curl_exec", dom, 1, date_str, commit=False)
            self.record_daily_transition(f"airlock:{dom}", f"exec:{dom}", 1, date_str, commit=False)
            return
        if not is_scanner and self._CLI_PIPE_UA_RE.search(ua_s) and clean_p in ("/", "/install.sh"):
            self.record_daily_metric("curl_airlock", dom, 1, date_str, commit=False)
            self.record_daily_transition(f"door:{dom}", f"airlock:{dom}", 1, date_str, commit=False)
            return
        if _STATIC_SUBRESOURCE_RE.search(raw_p):
            return
        if str(served_md) == "1":
            self.record_daily_metric("md_negotiated", dom, 1, date_str, commit=False)
        if self.is_likely_human_ua(ua_s):
            self.record_daily_metric("human_visit", dom, 1, date_str, commit=False)
            norm_ref = normalize_external_referrer(referrer, raw_p)
            if norm_ref is not None:
                ap = extract_registered_domain(norm_ref[1])
                if ap in ("youtube.com", "youtu.be"):
                    self.record_daily_metric("yt_referral_human", dom, 1, date_str, commit=False)
                    self.record_daily_transition("ref:youtube", f"visit:{dom}", 1, date_str, commit=False)
                elif ap in self._SOCIAL_IN_APEX:
                    self.record_daily_metric("social_in", ap, 1, date_str, commit=False)
                    self.record_daily_transition(f"ref:{ap}", f"visit:{dom}", 1, date_str, commit=False)
[[[DIVIDER]]]
    def _record_request_funnel_metrics(self, date_str, ip, ua, path, status, referrer, served_md, host, is_pirate, ip_id=None) -> None:
        """Update daily_metrics (N + P) and daily_transitions inside log_request's active transaction."""
        ip_s = str(ip or "")
        if ip_s.startswith(("127.", "10.", "192.168.")):
            return
        eff_ip = ip_id if ip_id is not None else self._get_or_create_id("ips", ip_s)
        def _rm(m: str, d: str) -> None:
            self.record_daily_metric(m, d, 1, date_str, ip_id=eff_ip, commit=False)
        dom = (host or "mikelev.in").strip().lower().rstrip(".").removeprefix("www.") or "mikelev.in"
        raw_p, st, ua_s = str(path or "/"), int(status or 0), str(ua or "")
        clean_p, ua_low = raw_p.split("?", 1)[0], ua_s.lower()
        if is_pirate or _PIRATE_TARGET_RE.search(clean_p):
            _rm("ids_pirate", "418" if st == 418 else f"{max(0, st) // 100}xx")
            return
        if 200 <= st < 400 and (clean_p == "/discord" or clean_p.startswith("/go/")):
            slug = "discord" if clean_p == "/discord" else (clean_p[4:].strip("/") or "root")[:40]
            _rm("social_out", slug)
            self.record_daily_transition(f"visit:{dom}", f"social:{slug}", 1, date_str, commit=False)
            return
        if not (200 <= st < 300):
            return
        if clean_p in ("/brain.zip", "/book.zip"):
            arch = clean_p.lstrip("/")
            _rm("archive_dl", arch)
            if self.is_likely_human_ua(ua_s):
                self.record_daily_transition(f"visit:{dom}", f"dl:{arch}", 1, date_str, commit=False)
            return
        is_scanner = any(b in ua_low for b in ("qlyzebot", "google-safety", "bot", "spider", "crawler"))
        if not is_scanner and (clean_p in ("/manifest.txt", "/key.rot") or "src=install" in raw_p or ua_low.startswith("sunpug-installer/")):
            _rm("curl_exec", dom)
            self.record_daily_transition(f"airlock:{dom}", f"exec:{dom}", 1, date_str, commit=False)
            return
        if not is_scanner and self._CLI_PIPE_UA_RE.search(ua_s) and clean_p in ("/", "/install.sh"):
            _rm("curl_airlock", dom)
            self.record_daily_transition(f"door:{dom}", f"airlock:{dom}", 1, date_str, commit=False)
            return
        if _STATIC_SUBRESOURCE_RE.search(raw_p):
            return
        if str(served_md) == "1":
            _rm("md_negotiated", dom)
        if self.is_likely_human_ua(ua_s):
            _rm("human_visit", dom)
            norm_ref = normalize_external_referrer(referrer, raw_p)
            if norm_ref is not None:
                ap = extract_registered_domain(norm_ref[1])
                if ap in ("youtube.com", "youtu.be"):
                    _rm("yt_referral_human", dom)
                    self.record_daily_transition("ref:youtube", f"visit:{dom}", 1, date_str, commit=False)
                elif ap in self._SOCIAL_IN_APEX:
                    _rm("social_in", ap)
                    self.record_daily_transition(f"ref:{ap}", f"visit:{dom}", 1, date_str, commit=False)
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
            queries = (
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'ids_pirate', CASE WHEN status = 418 THEN '418' ELSE CAST(status / 100 AS TEXT) || 'xx' END, date, SUM(count) FROM pirate_logs GROUP BY 1, 2, 3",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'md_negotiated', 'mikelev.in', t.date, SUM(t.count) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN ltrim(p.value, '/') WHEN p.value = '/discord' THEN 'discord' ELSE substr(p.value, 5) END, l.date, SUM(l.count) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2, 3",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_airlock', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%') GROUP BY d.value, ddl.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_exec', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND (p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%' GROUP BY d.value, ddl.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN d.value ELSE rdl.ref_apex END, rdl.date, SUM(rdl.human_count) FROM referrer_daily_logs rdl JOIN domains d ON rdl.domain_id = d.id WHERE rdl.human_count > 0 AND rdl.ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2, 3",
            )
[[[DIVIDER]]]
            queries = (
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'ids_pirate', CASE WHEN status = 418 THEN '418' ELSE CAST(status / 100 AS TEXT) || 'xx' END, date, SUM(count), COUNT(DISTINCT ip_id) FROM pirate_logs GROUP BY 1, 2, 3",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'md_negotiated', 'mikelev.in', t.date, SUM(t.count), COUNT(DISTINCT t.ip_id) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN ltrim(p.value, '/') WHEN p.value = '/discord' THEN 'discord' ELSE substr(p.value, 5) END, l.date, SUM(l.count), COUNT(DISTINCT l.ip_id) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2, 3",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'curl_airlock', d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%') GROUP BY d.value, ddl.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'curl_exec', d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND (p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%' GROUP BY d.value, ddl.date",
                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN d.value ELSE rdl.ref_apex END, rdl.date, SUM(rdl.human_count), SUM(rdl.unique_ips) FROM referrer_daily_logs rdl JOIN domains d ON rdl.domain_id = d.id WHERE rdl.human_count > 0 AND rdl.ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2, 3",
            )
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
    def get_daily_series(self, metric: str, dim: str = "all", days: int = 14, end_date: str | None = None) -> list[tuple[str, int]]:
        """Return a zero-filled contiguous [(YYYY-MM-DD, count), ...] daily series for honest OLS slopes."""
        end_d = datetime.date.fromisoformat(end_date) if end_date else datetime.date.today()
        n_days = max(1, int(days))
        start_d = end_d - datetime.timedelta(days=n_days - 1)
        span = [(start_d + datetime.timedelta(days=i)).isoformat() for i in range(n_days)]
        rows = self.get_conn().execute("SELECT day, COALESCE(SUM(count), 0) FROM daily_metrics WHERE metric = ? AND (? = '*' OR dim = ?) AND day BETWEEN ? AND ? GROUP BY day", (str(metric), str(dim), str(dim), span[0], span[-1])).fetchall()
        by_day = {str(d): int(c or 0) for d, c in rows}
        return [(d, by_day.get(d, 0)) for d in span]
[[[DIVIDER]]]
    def get_daily_series(self, metric: str, dim: str = "all", days: int = 14, end_date: str | None = None, unique_ips: bool = False) -> list[tuple[str, int]]:
        """Return a zero-filled contiguous [(YYYY-MM-DD, value), ...] daily series for N hits or P unique_ips."""
        end_d = datetime.date.fromisoformat(end_date) if end_date else datetime.date.today()
        n_days = max(1, int(days))
        start_d = end_d - datetime.timedelta(days=n_days - 1)
        span = [(start_d + datetime.timedelta(days=i)).isoformat() for i in range(n_days)]
        col = "unique_ips" if unique_ips else "count"
        rows = self.get_conn().execute(f"SELECT day, COALESCE(SUM({col}), 0) FROM daily_metrics WHERE metric = ? AND (? = '*' OR dim = ?) AND day BETWEEN ? AND ? GROUP BY day", (str(metric), str(dim), str(dim), span[0], span[-1])).fetchall()
        by_day = {str(d): int(c or 0) for d, c in rows}
        return [(d, by_day.get(d, 0)) for d in span]
[[[REPLACE]]]
Target: remotes/honeybot/scripts/db.py
[[[SEARCH]]]
    def get_curl_airlock_split(self, door: str = "sunpug.io", days: int = 14, end_date: str | None = None) -> dict:
        """Separate Stage-1 curl|less inspection from Stage-2 curl|bash execution for a Sun Pug door (or '*')."""
        s1 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date)
        s2 = self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date)
        return {"door": door, "airlock": s1, "exec": s2, "inspect_only": [(d, max(0, a - e)) for (d, a), (_, e) in zip(s1, s2)]}
[[[DIVIDER]]]
    def get_curl_airlock_split(self, door: str = "sunpug.io", days: int = 14, end_date: str | None = None) -> dict:
        """Separate Stage-1 curl|less inspection from Stage-2 curl|bash execution for a Sun Pug door (or '*')."""
        s1, s2 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date), self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date)
        p1, p2 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date, unique_ips=True), self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date, unique_ips=True)
        return {"door": door, "airlock": s1, "exec": s2, "inspect_only": [(d, max(0, a - e)) for (d, a), (_, e) in zip(s1, s2)], "airlock_ips": p1, "exec_ips": p2}
[[[REPLACE]]]
```

**Car 3b — Add `--funnel` `(N, P)` daily OLS sparklines, per-door `curl | less` vs. `curl | bash` split, `river` next-day forecasts, and Markov next-step prescriptions to `scripts/markov.py` (`112/220` lines):**

```text
Target: scripts/markov.py
[[[SEARCH]]]
from river import drift, linear_model, optim, preprocessing, stats
[[[DIVIDER]]]
from river import drift, linear_model, optim, preprocessing, stats

_REPO_ROOT = Path(__file__).resolve().parents[1]
if str(_REPO_ROOT) not in sys.path:
    sys.path.insert(0, str(_REPO_ROOT))
from imports.ascii_displays import compute_ols_velocity, render_sparkline
[[[REPLACE]]]
Target: scripts/markov.py
[[[SEARCH]]]
def run_demo_series() -> int:
    synth_num = [
        f"t{i} {100.0 + 0.8 * i + (3.5 if i % 4 in (0, 1) else -2.0):.2f}"
        for i in range(1, 25)
    ]
    return analyze_generic_series(synth_num, label="demo-quant-series", top_k=4)
[[[DIVIDER]]]
_REMOTE_FUNNEL_PY = r"""
import datetime, json, sqlite3, sys
days = max(2, int(sys.argv[1]))
db_p = sys.argv[2] if len(sys.argv) > 2 and sys.argv[2] else "/home/mike/www/mikelev.in/honeybot.db"
conn = sqlite3.connect(f"file:{db_p}?mode=ro", uri=True)
end_d = datetime.date.today()
span = [(end_d - datetime.timedelta(days=days - 1 - i)).isoformat() for i in range(days)]
d0, d1, idx_m = span[0], span[-1], {d: i for i, d in enumerate(span)}
tbls = {r[0] for r in conn.execute("SELECT name FROM sqlite_master WHERE type='table'").fetchall()}
metrics, doors, trans = {}, {}, []
def _bump(k, dy, n_v, p_v):
    if str(dy) in idx_m:
        a = metrics.setdefault(k, [[0, 0] for _ in range(days)])[idx_m[str(dy)]]
        a[0], a[1] = a[0] + int(n_v or 0), a[1] + int(p_v or 0)
def _door(dom, kind, dy, n_v, p_v):
    if dom and str(dy) in idx_m:
        r = doors.setdefault(dom, {x: [0] * days for x in ("air_n", "air_p", "exe_n", "exe_p")})
        r[f"{kind}_n"][idx_m[str(dy)]] += int(n_v or 0)
        r[f"{kind}_p"][idx_m[str(dy)]] += int(p_v or 0)
has_dm = "daily_metrics" in tbls and (conn.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0] > 0
if has_dm:
    dm_cols = {r[1] for r in conn.execute("PRAGMA table_info(daily_metrics)").fetchall()}
    p_col = "unique_ips" if "unique_ips" in dm_cols else "count"
    for m, dm, dy, c, u in conn.execute(f"SELECT metric, dim, day, count, {p_col} FROM daily_metrics WHERE day BETWEEN ? AND ?", (d0, d1)):
        _bump(m, dy, c, u)
        if m in ("curl_airlock", "curl_exec"):
            _door(dm, "air" if m == "curl_airlock" else "exe", dy, c, u)
    if "daily_transitions" in tbls:
        trans = [[f, t, int(c)] for f, t, c in conn.execute("SELECT from_state, to_state, SUM(count) FROM daily_transitions WHERE day BETWEEN ? AND ? GROUP BY 1, 2 ORDER BY 3 DESC LIMIT 20", (d0, d1))]
    mode = "live_metrics"
else:
    if "pirate_logs" in tbls:
        [_bump("ids_pirate", dy, c, u) for dy, c, u in conn.execute("SELECT date, SUM(count), COUNT(DISTINCT ip_id) FROM pirate_logs WHERE date BETWEEN ? AND ? GROUP BY date", (d0, d1))]
    if "telemetry" in tbls:
        [_bump("md_negotiated", dy, c, u) for dy, c, u in conn.execute("SELECT t.date, SUM(t.count), COUNT(DISTINCT t.ip_id) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND t.date BETWEEN ? AND ? AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date", (d0, d1))]
    if "daily_logs" in tbls:
        [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, l.date, SUM(l.count), COUNT(DISTINCT l.ip_id) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE l.date BETWEEN ? AND ? AND ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2", (d0, d1))]
    if "domain_daily_logs" in tbls:
        for knd, dom, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/', '/install.sh') THEN 'air' ELSE 'exe' END, d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.date BETWEEN ? AND ? AND ddl.status BETWEEN 200 AND 299 AND ((p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%')) OR ((p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%')) GROUP BY 1, 2, 3", (d0, d1)):
            _bump("curl_airlock" if knd == "air" else "curl_exec", dy, c, u)
            _door(dom, knd, dy, c, u)
    if "referrer_daily_logs" in tbls:
        [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, date, SUM(human_count), SUM(unique_ips) FROM referrer_daily_logs WHERE date BETWEEN ? AND ? AND human_count > 0 AND ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2", (d0, d1))]
    mode = "live_derived"
print(json.dumps({"status": mode, "span": span, "metrics": metrics, "doors": doors, "transitions": trans}))
"""

def _river_next_day(vals: list[int]) -> float:
    model = preprocessing.StandardScaler() | linear_model.LinearRegression(optimizer=optim.SGD(0.02))
    ew = stats.EWMean(fading_factor=0.35)
    for i, v in enumerate(vals):
        if i > 0:
            model.learn_one({"t": float(i), "prev": float(vals[i - 1]), "ewma": float(ew.get() or v)}, float(v))
        ew.update(float(v))
    return max(0.0, float(model.predict_one({"t": float(len(vals)), "prev": float(vals[-1]), "ewma": float(ew.get() or vals[-1])}))) if len(vals) >= 2 else float(vals[-1] if vals else 0.0)

def run_funnel_compass(days: int = 14, top_k: int = 5) -> int:
    db_ov, host = os.environ.get("PIPULATE_HONEYBOT_DB", "").strip(), os.environ.get("PIPULATE_HONEYBOT_HOST", "honeybot")
    cmd = [sys.executable, "-", str(int(days)), db_ov] if (db_ov and Path(db_ov).is_file()) else ["ssh", "-o", "BatchMode=yes", "-o", "ConnectTimeout=3", host, "python3", "-", str(int(days))]
    res = subprocess.run(cmd, input=_REMOTE_FUNNEL_PY, capture_output=True, text=True, timeout=6, check=False)
    snap = json.loads(res.stdout.strip()) if res.returncode == 0 and (res.stdout or "").strip() else {"status": "offline", "span": [], "metrics": {}, "doors": {}, "transitions": []}
    span = snap.get("span") or []
    n_d, m_map = len(span) or max(2, int(days)), snap.get("metrics") or {}
    def _np(key: str) -> tuple[list[int], list[int]]:
        arr = m_map.get(key) or [[0, 0] for _ in range(n_d)]
        return [int(x[0]) for x in arr], [int(x[1]) for x in arr]
    air_n, air_p = _np("curl_airlock")
    exe_n, exe_p = _np("curl_exec")
    less_n, less_p = [max(0, a - e) for a, e in zip(air_n, exe_n)], [max(0, a - e) for a, e in zip(air_p, exe_p)]
    tracks = [("archive_dl", *_np("archive_dl")), ("curl_less_inspect", less_n, less_p), ("curl_bash_exec", exe_n, exe_p), ("yt_referral_human", *_np("yt_referral_human")), ("social_in", *_np("social_in")), ("social_out", *_np("social_out")), ("md_negotiated", *_np("md_negotiated")), ("ids_pirate", *_np("ids_pirate"))]
    print(f"# 🧭 PIPULATE DIRECTIONAL VELOCITY & RIVER FUNNEL ({n_d}d: {span[0] + '..' + span[-1] if span else n_d} | status={snap.get('status')})")
    print("\n## 1. Daily Directional Counters (Sparkline, OLS Slope, N Hits / P Daily IPs, River t+1)")
    slopes = {}
    for name, nv, pv in tracks:
        st = slopes[name] = compute_ols_velocity(nv)
        print(f"  • {name:<18s} {(render_sparkline(nv, width=14) or '—'):<14s} {st['arrow']} {st['slope']:+6.2f}/d | N={int(st['total']):5d}, P={sum(pv):4d} IPs | river_t+1={_river_next_day(nv):5.1f}")
    print("\n## 2. Sun Pug Per-Door Airlock Split (curl | less vs. curl | bash)")
    doors = snap.get("doors") or {}
    for dk in sorted(doors, key=lambda k: sum(doors[k]["air_n"]) + sum(doors[k]["exe_n"]), reverse=True)[: max(4, top_k)]:
        dn, dp, en = doors[dk]["air_n"], doors[dk]["air_p"], doors[dk]["exe_n"]
        st, tot_a, tot_e = compute_ols_velocity(dn), sum(dn), sum(en)
        print(f"  • {dk:<18s} {render_sparkline(dn, width=14):<14s} {st['arrow']} {st['slope']:+6.2f}/d | less={sum(max(0, a - e) for a, e in zip(dn, en)):4d}, bash={tot_e:4d} ({(100.0 * tot_e / tot_a if tot_a else 0.0):3.0f}% exec, P={sum(dp):3d} IPs)")
    if not doors:
        print("  • (No curl/Wget door hits recorded in window yet)")
    print("\n## 3. Markov Transition Odds & What To Do Next")
    f_tot = Counter()
    for f_s, _, cnt in (snap.get("transitions") or []):
        f_tot[f_s] += int(cnt)
    for f_s, t_s, cnt in (snap.get("transitions") or [])[:top_k]:
        print(f"  • P({t_s} | {f_s}) = {(100.0 * int(cnt) / max(1, f_tot[f_s])):5.1f}% (n={cnt})")
    acts = []
    if sum(less_n) > sum(exe_n) and sum(less_n) >= 1:
        acts.append("Inspect-only `curl | less` leads `curl | bash` -> wire Stage-2 `/manifest.txt` check in `install.sh` (P0.61/P0.78).")
    acts.append(f"`brain.zip` slope is {slopes['archive_dl']['slope']:+.2f}/d -> {'keep `/book.zip` synced on publish (P0.67).' if slopes['archive_dl']['slope'] > 0 else 'spotlight `brain.zip` + WOPR hook in next Norma cone.'}")
    if slopes["ids_pirate"]["total"] > 0:
        acts.append(f"Pirate 4xx/418 scanner pressure active (N={int(slopes['ids_pirate']['total'])}) -> prime HTTP 418 Rickroll Weebo show (P0.80).")
    for idx, act in enumerate(acts[:3], 1):
        print(f"  {idx}. NEXT: {act}")
    return 0

def run_demo_series() -> int:
    synth_num = [
        f"t{i} {100.0 + 0.8 * i + (3.5 if i % 4 in (0, 1) else -2.0):.2f}"
        for i in range(1, 25)
    ]
    return analyze_generic_series(synth_num, label="demo-quant-series", top_k=4)
[[[REPLACE]]]
Target: scripts/markov.py
[[[SEARCH]]]
    parser.add_argument("--days", type=int, default=30, help="Lookback window in days for --radar (default: 30).")
    parser.add_argument("--stdin", action="store_true", help="Read series observations from stdin.")
    parser.add_argument("--demo-series", action="store_true", help="Run built-in numeric time-series verification.")
    args = parser.parse_args()
    if args.radar or args.ai or args.links or args.by_domain:
[[[DIVIDER]]]
    parser.add_argument("--days", type=int, default=30, help="Lookback window in days for --radar / --funnel (default: 30).")
    parser.add_argument("--funnel", action="store_true", help="Show daily directional sparklines (N/P), OLS slopes, & River/Markov funnel.")
    parser.add_argument("--stdin", action="store_true", help="Read series observations from stdin.")
    parser.add_argument("--demo-series", action="store_true", help="Run built-in numeric time-series verification.")
    args = parser.parse_args()
    if args.funnel:
        return run_funnel_compass(days=14 if "--days" not in sys.argv else args.days, top_k=args.top)
    if args.radar or args.ai or args.links or args.by_domain:
[[[REPLACE]]]
```

**Car 3c — Expose reusable `Pipulate.sparkline()` and `Pipulate.sparkline_card()` FastHTML/Notebook helpers in `pipulate/core.py` (`28/220` lines):**

```text
Target: pipulate/core.py
[[[SEARCH]]]
        from imports import ascii_displays as aa
        return aa.patronus(name, duration=duration)
[[[DIVIDER]]]
        from imports import ascii_displays as aa
        return aa.patronus(name, duration=duration)

    def sparkline(self, values, width: int = 128, height: int = 28, stroke: str = "#34d399", show_ols: bool = True, as_text: bool = False):
        """Render a numeric series as a zero-JS FastHTML inline SVG sparkline (with OLS trendline) or Unicode text."""
        from fasthtml.common import NotStr
        from imports import ascii_displays as aa
        return aa.render_sparkline(values) if as_text else NotStr(aa.render_sparkline_svg(values, width=width, height=height, stroke=stroke, show_ols=show_ols))

    def sparkline_card(self, title: str, values, subtitle: str = "", unit: str = "/d", unique_values=None, width: int = 136, height: int = 30):
        """Render a PicoCSS FastHTML Card pairing an inline SVG sparkline with OLS velocity slope and (N, P) totals."""
        from fasthtml.common import Div, Small, Span, Strong
        from imports import ascii_displays as aa
        st = aa.compute_ols_velocity(values)
        stroke = "#34d399" if st["slope"] > 0.02 else ("#f87171" if st["slope"] < -0.02 else "#60a5fa")
        p_str = f" · P={int(round(sum(aa._coerce_numeric_series(unique_values))))} IPs" if unique_values is not None else ""
        meta = subtitle or f"N={int(round(st['total']))}{p_str} · last={int(round(st['latest']))} · R²={st['r2']:.2f}"
        return Card(
            Div(
                Div(Strong(title), Small(meta, style="display:block;opacity:0.78;font-family:monospace;font-size:0.78rem;"), style="min-width:0;"),
                Div(
                    self.sparkline(values, width=width, height=height, stroke=stroke, show_ols=True),
                    Span(f"{st['arrow']} {st['slope']:+.2f}{unit}", style=f"font-family:monospace;font-weight:700;font-size:0.86rem;color:{stroke};white-space:nowrap;"),
                    style="display:flex;align-items:center;gap:0.65rem;flex-shrink:0;",
                ),
                style="display:flex;align-items:center;justify-content:space-between;gap:1rem;flex-wrap:wrap;",
            ),
            style="padding:0.7rem 0.95rem;margin-bottom:0.65rem;",
        )
[[[REPLACE]]]
```

Ignition: none (all four probes load `db.py`, `scripts/markov.py`, and `pipulate/core.py` directly at compile time, and `_REMOTE_FUNNEL_PY` derives live `(N, P)` sparklines from Honeybot in read-only mode even before `nixops.sh` syncs `db.py`).

### (4) PROMPT

```text
Verify the AFTER receipts for `DB_NP_GATE_STRADDLE` (`has_unique_ips=True`, `np_dedupe_and_purge_ok=True`, `stale_gate=0`), `MARKOV_FUNNEL_STRADDLE` (`has_flag=True`, `funnel_ok=True`), `PIP_SPARKLINE_WIDGET_STRADDLE` (`has_methods=True`, `widget_ok=True`, `ruff_rc=0`), and `LIVE_HONEYBOT_FUNNEL` (live 14-day `(N, P)` sparklines, per-door Sun Pug split, and `river` forecasts from Honeybot). Assuming all four are green, inspect `apps/220_roadmap.py` and `remotes/honeybot/scripts/forest.py` to wire live directional sparklines + OLS velocity slopes into a real FastHTML app surface (satisfying `P0.79`'s `"Gate: sparkline + slope render in app"`) and Honeybot's report vocabulary so we are ready to sync Honeybot and dismount cleanly.
```

### (5) EXTERNAL DELIVERABLES

No external deliverables.

**MikeLev.in**: 



## THE AI-EDIT METHOD

> Same commands, run twice, one change between them. Where the readings
> differ is what the change did; the diff in the middle is the receipt.

**1: BEFORE** (PROBE): 

```bash
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst = hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok} n={n_s} p={p_s} stale_gate={old_g}")'
.venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7, day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")'
.venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok} ruff_rc={r_rc}")'
.venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out = (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")'
DB_NP_GATE_STRADDLE: has_unique_ips=False np_dedupe_and_purge_ok=False n=[] p=[] stale_gate=-1
MARKOV_FUNNEL_STRADDLE: has_flag=False funnel_ok=False lines=0
PIP_SPARKLINE_WIDGET_STRADDLE: has_methods=False widget_ok=False ruff_rc=0
LIVE_HONEYBOT_FUNNEL:
FUNNEL_PENDING
(nix) pipulate $ 
```

**2: AFTER** (NEXT CONTEXT): 

```text
# # Context 1: The no-traditional opening
# context.txt
# ! book
# ! rgxc 40
# ! markov
# ! latest

# # Context 2: Bank IDS/Plug TODOs and inspect db.py + markov.py for daily OLS sparklines & River funnel
# context.txt
# foo_files.py
# remotes/honeybot/scripts/db.py
# scripts/markov.py
# imports/ascii_displays.py
# ! .venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
# ! .venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
# ! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'

# # Context 3: Wire markov.py --funnel River predictor and Pipulate core.py sparkline helper
# context.txt
# ! markov
# scripts/markov.py
# remotes/honeybot/scripts/db.py
# imports/ascii_displays.py
# pipulate/__init__.py
# pipulate/core.py
# ! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); has_nota = "Notaplebius:" in foo_txt; has_spore = "CORDYCEPS" in foo_txt; p0_ok = all(f"P0.{n}" in foo_txt for n in (78, 79, 80, 81)); print(f"FOO_NOTAPLEBIUS_CHECK: notaplebius={has_nota} cordyceps_todo={has_spore} p0_78_81={p0_ok}")'
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; inst = hdb.HoneyDB(":memory:"); has_fn = hasattr(inst, "get_daily_series") and hasattr(inst, "get_curl_airlock_split"); _ = (inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "sunpug-installer/1.0", "/manifest.txt", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "python-requests/2.31", "/wp-login.php", 418, date_str="2026-11-21", host="mikelev.in", is_pirate=True)) if has_fn else None; dl_s = inst.get_daily_series("archive_dl", "brain.zip", days=3, end_date="2026-11-21") if has_fn else []; sp_d = inst.get_curl_airlock_split("sunpug.io", days=2, end_date="2026-11-21") if has_fn else {}; s_ok = (dl_s == [("2026-11-19", 0), ("2026-11-20", 0), ("2026-11-21", 1)]); split_ok = (sp_d.get("inspect_only") == [("2026-11-20", 1), ("2026-11-21", 0)]); print(f"DB_FUNNEL_STRADDLE: has_fn={has_fn} zero_filled_dl={s_ok} curl_less_split={split_ok}")'
# ! .venv/bin/python scripts/pyprobe.py 'from imports import ascii_displays as ad; has_ols = hasattr(ad, "compute_ols_velocity") and hasattr(ad, "render_sparkline") and hasattr(ad, "render_sparkline_svg"); up = ad.compute_ols_velocity([2, 4, 6, 8, 10]) if has_ols else {}; arr = up.get("arrow"); z_spk = ad.render_sparkline([0, 0, 0, 0]) if has_ols else ""; svg = ad.render_sparkline_svg([1, 3, 2, 5]) if has_ols else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "imports/ascii_displays.py"); ols_ok = bool(has_ols and abs(up.get("slope", 0.0) - 2.0) < 1e-9 and abs(up.get("r2", 0.0) - 1.0) < 1e-9 and z_spk == "▁▁▁▁" and "<polyline" in svg and "<line" in svg); print(f"OLS_SPARKLINE_STRADDLE: has_ols={has_ols} ols_math_ok={ols_ok} arrow={arr} ruff_rc={r_rc}")'

# Context 4: Verify live Honeybot (N, P) funnel compass and wire sparkline cards into FastHTML app
context.txt
! markov
apps/220_roadmap.py
remotes/honeybot/scripts/forest.py
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst = hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok} n={n_s} p={p_s} stale_gate={old_g}")'
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7, day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")'
! .venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out = (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")'
```

**3: CHANGE** (PATCHES): 

```diff
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'remotes/honeybot/scripts/db.py'.
📏 Car meter: 1/5 file(s), 8/12 block(s), 100/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/remotes/honeybot/scripts/db.py b/remotes/honeybot/scripts/db.py
index dce0847..83ba688 100644
--- a/remotes/honeybot/scripts/db.py
+++ b/remotes/honeybot/scripts/db.py
@@ -283,14 +283,19 @@ class HoneyDB:
         cur.execute("CREATE INDEX IF NOT EXISTS idx_pirate_logs_ip_id ON pirate_logs(ip_id)")
         cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_radar_seen ON referrer_radar(first_seen DESC)")
         cur.execute("CREATE INDEX IF NOT EXISTS idx_referrer_daily_date_path ON referrer_daily_logs(date, path_id)")
-        cur.execute("CREATE TABLE IF NOT EXISTS daily_metrics (metric TEXT NOT NULL, dim TEXT NOT NULL DEFAULT 'all', day TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (metric, dim, day)) WITHOUT ROWID")
+        dm_cols = {r[1] for r in cur.execute("PRAGMA table_info(daily_metrics)").fetchall()}
+        need_dm_rebackfill = bool(dm_cols and "unique_ips" not in dm_cols)
+        if need_dm_rebackfill:
+            cur.execute("ALTER TABLE daily_metrics ADD COLUMN unique_ips INTEGER NOT NULL DEFAULT 0")
+        cur.execute("CREATE TABLE IF NOT EXISTS daily_metrics (metric TEXT NOT NULL, dim TEXT NOT NULL DEFAULT 'all', day TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, unique_ips INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (metric, dim, day)) WITHOUT ROWID")
+        cur.execute("CREATE TABLE IF NOT EXISTS daily_metric_ip_gate (day TEXT NOT NULL, metric TEXT NOT NULL, dim TEXT NOT NULL, ip_id INTEGER NOT NULL, PRIMARY KEY (day, metric, dim, ip_id)) WITHOUT ROWID")
         cur.execute("CREATE TABLE IF NOT EXISTS daily_transitions (day TEXT NOT NULL, from_state TEXT NOT NULL, to_state TEXT NOT NULL, count INTEGER NOT NULL DEFAULT 0, last_updated TIMESTAMP DEFAULT CURRENT_TIMESTAMP, PRIMARY KEY (day, from_state, to_state)) WITHOUT ROWID")
 
         conn.commit()
         self.backfill_rfc9421_from_access_log()
         self.backfill_dns_audit()
         self.backfill_external_referrers()
-        self.rollup_daily_metrics()
+        self.rollup_daily_metrics(force=need_dm_rebackfill)
 
     def backfill_external_referrers(self):
         """One-time idempotent backfill of external_referrers, referrer_radar, & referrer_daily_logs."""
@@ -545,7 +550,7 @@ class HoneyDB:
                 conn.execute("INSERT INTO referrer_radar (ref_apex, domain_id, path_id, first_referrer, last_referrer, total_count, human_count, bot_count) VALUES (?, ?, ?, ?, ?, 1, ?, ?) ON CONFLICT(ref_apex, domain_id, path_id) DO UPDATE SET last_referrer = excluded.last_referrer, last_seen = CURRENT_TIMESTAMP, total_count = total_count + 1, human_count = human_count + excluded.human_count, bot_count = bot_count + excluded.bot_count", (ref_apex, dom_id_r, r_pid, canon_ref, canon_ref, is_hum, 1 - is_hum))
                 conn.execute("INSERT INTO referrer_daily_logs (date, ref_apex, domain_id, path_id, sample_referrer, count, human_count, bot_count, unique_ips) VALUES (?, ?, ?, ?, ?, 1, ?, ?, ?) ON CONFLICT(date, ref_apex, domain_id, path_id) DO UPDATE SET sample_referrer = CASE WHEN length(excluded.sample_referrer) > length(referrer_daily_logs.sample_referrer) THEN excluded.sample_referrer ELSE referrer_daily_logs.sample_referrer END, count = count + 1, human_count = human_count + excluded.human_count, bot_count = bot_count + excluded.bot_count, unique_ips = unique_ips + excluded.unique_ips, last_updated = CURRENT_TIMESTAMP", (date_str, ref_apex, dom_id_r, r_pid, canon_ref, is_hum, 1 - is_hum, new_ip))
 
-        self._record_request_funnel_metrics(date_str, ip, ua, path, status, referrer, served_md, host, is_pirate)
+        self._record_request_funnel_metrics(date_str, ip, ua, path, status, referrer, served_md, host, is_pirate, ip_id=ip_id)
         conn.commit()
         if ua and ("Googlebot" in ua or "bingbot" in ua.lower() or "Applebot" in ua):
             self.verify_fcrdns_ip(ip, ua)
@@ -955,11 +960,18 @@ class HoneyDB:
     _CLI_PIPE_UA_RE = re.compile(r"^(?:curl|Wget|fetch\s+libfetch|sunpug-installer)/", re.IGNORECASE)
     _SOCIAL_IN_APEX = frozenset({"t.co", "x.com", "twitter.com", "linkedin.com", "lnkd.in", "reddit.com", "facebook.com", "bsky.app", "news.ycombinator.com", "lobste.rs"})
 
-    def record_daily_metric(self, metric: str, dim: str = "all", amount: int = 1, day: str | None = None, commit: bool = True) -> None:
-        """Increment a compact daily directional counter (1 row per metric + dim + UTC day)."""
+    def record_daily_metric(self, metric: str, dim: str = "all", amount: int = 1, day: str | None = None, ip_id: int | None = None, commit: bool = True) -> None:
+        """Increment a compact daily directional counter (N hits + P self-erasing 24h unique IPs)."""
         d_str = day or datetime.date.today().isoformat()
+        m_s, dm_s = str(metric), str(dim or "all")
         conn = self.get_conn()
-        conn.execute("INSERT INTO daily_metrics (metric, dim, day, count) VALUES (?, ?, ?, ?) ON CONFLICT(metric, dim, day) DO UPDATE SET count = count + excluded.count, last_updated = CURRENT_TIMESTAMP", (str(metric), str(dim or "all"), d_str, int(amount)))
+        new_ip = 0
+        if ip_id is not None:
+            if getattr(self, "_metric_ip_gate_date", None) != d_str:
+                conn.execute("DELETE FROM daily_metric_ip_gate WHERE day < ?", (d_str,))
+                self._metric_ip_gate_date = d_str
+            new_ip = 1 if conn.execute("INSERT OR IGNORE INTO daily_metric_ip_gate VALUES (?, ?, ?, ?)", (d_str, m_s, dm_s, int(ip_id))).rowcount > 0 else 0
+        conn.execute("INSERT INTO daily_metrics (metric, dim, day, count, unique_ips) VALUES (?, ?, ?, ?, ?) ON CONFLICT(metric, dim, day) DO UPDATE SET count = count + excluded.count, unique_ips = unique_ips + excluded.unique_ips, last_updated = CURRENT_TIMESTAMP", (m_s, dm_s, d_str, int(amount), new_ip))
         if commit:
             conn.commit()
 
@@ -971,56 +983,56 @@ class HoneyDB:
         if commit:
             conn.commit()
 
-    def _record_request_funnel_metrics(self, date_str, ip, ua, path, status, referrer, served_md, host, is_pirate) -> None:
-        """Update daily_metrics and daily_transitions inside log_request's active transaction."""
+    def _record_request_funnel_metrics(self, date_str, ip, ua, path, status, referrer, served_md, host, is_pirate, ip_id=None) -> None:
+        """Update daily_metrics (N + P) and daily_transitions inside log_request's active transaction."""
         ip_s = str(ip or "")
         if ip_s.startswith(("127.", "10.", "192.168.")):
             return
+        eff_ip = ip_id if ip_id is not None else self._get_or_create_id("ips", ip_s)
+        def _rm(m: str, d: str) -> None:
+            self.record_daily_metric(m, d, 1, date_str, ip_id=eff_ip, commit=False)
         dom = (host or "mikelev.in").strip().lower().rstrip(".").removeprefix("www.") or "mikelev.in"
-        raw_p = str(path or "/")
-        clean_p = raw_p.split("?", 1)[0]
-        st = int(status or 0)
-        ua_s = str(ua or "")
-        ua_low = ua_s.lower()
+        raw_p, st, ua_s = str(path or "/"), int(status or 0), str(ua or "")
+        clean_p, ua_low = raw_p.split("?", 1)[0], ua_s.lower()
         if is_pirate or _PIRATE_TARGET_RE.search(clean_p):
-            self.record_daily_metric("ids_pirate", "418" if st == 418 else f"{max(0, st) // 100}xx", 1, date_str, commit=False)
+            _rm("ids_pirate", "418" if st == 418 else f"{max(0, st) // 100}xx")
             return
         if 200 <= st < 400 and (clean_p == "/discord" or clean_p.startswith("/go/")):
             slug = "discord" if clean_p == "/discord" else (clean_p[4:].strip("/") or "root")[:40]
-            self.record_daily_metric("social_out", slug, 1, date_str, commit=False)
+            _rm("social_out", slug)
             self.record_daily_transition(f"visit:{dom}", f"social:{slug}", 1, date_str, commit=False)
             return
         if not (200 <= st < 300):
             return
         if clean_p in ("/brain.zip", "/book.zip"):
             arch = clean_p.lstrip("/")
-            self.record_daily_metric("archive_dl", arch, 1, date_str, commit=False)
+            _rm("archive_dl", arch)
             if self.is_likely_human_ua(ua_s):
                 self.record_daily_transition(f"visit:{dom}", f"dl:{arch}", 1, date_str, commit=False)
             return
         is_scanner = any(b in ua_low for b in ("qlyzebot", "google-safety", "bot", "spider", "crawler"))
         if not is_scanner and (clean_p in ("/manifest.txt", "/key.rot") or "src=install" in raw_p or ua_low.startswith("sunpug-installer/")):
-            self.record_daily_metric("curl_exec", dom, 1, date_str, commit=False)
+            _rm("curl_exec", dom)
             self.record_daily_transition(f"airlock:{dom}", f"exec:{dom}", 1, date_str, commit=False)
             return
         if not is_scanner and self._CLI_PIPE_UA_RE.search(ua_s) and clean_p in ("/", "/install.sh"):
-            self.record_daily_metric("curl_airlock", dom, 1, date_str, commit=False)
+            _rm("curl_airlock", dom)
             self.record_daily_transition(f"door:{dom}", f"airlock:{dom}", 1, date_str, commit=False)
             return
         if _STATIC_SUBRESOURCE_RE.search(raw_p):
             return
         if str(served_md) == "1":
-            self.record_daily_metric("md_negotiated", dom, 1, date_str, commit=False)
+            _rm("md_negotiated", dom)
         if self.is_likely_human_ua(ua_s):
-            self.record_daily_metric("human_visit", dom, 1, date_str, commit=False)
+            _rm("human_visit", dom)
             norm_ref = normalize_external_referrer(referrer, raw_p)
             if norm_ref is not None:
                 ap = extract_registered_domain(norm_ref[1])
                 if ap in ("youtube.com", "youtu.be"):
-                    self.record_daily_metric("yt_referral_human", dom, 1, date_str, commit=False)
+                    _rm("yt_referral_human", dom)
                     self.record_daily_transition("ref:youtube", f"visit:{dom}", 1, date_str, commit=False)
                 elif ap in self._SOCIAL_IN_APEX:
-                    self.record_daily_metric("social_in", ap, 1, date_str, commit=False)
+                    _rm("social_in", ap)
                     self.record_daily_transition(f"ref:{ap}", f"visit:{dom}", 1, date_str, commit=False)
 
     def rollup_daily_metrics(self, force: bool = False) -> int:
@@ -1031,12 +1043,12 @@ class HoneyDB:
             if not force and (cur.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0] > 0:
                 return 0
             queries = (
-                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'ids_pirate', CASE WHEN status = 418 THEN '418' ELSE CAST(status / 100 AS TEXT) || 'xx' END, date, SUM(count) FROM pirate_logs GROUP BY 1, 2, 3",
-                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'md_negotiated', 'mikelev.in', t.date, SUM(t.count) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date",
-                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN ltrim(p.value, '/') WHEN p.value = '/discord' THEN 'discord' ELSE substr(p.value, 5) END, l.date, SUM(l.count) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2, 3",
-                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_airlock', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%') GROUP BY d.value, ddl.date",
-                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT 'curl_exec', d.value, ddl.date, SUM(ddl.count) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND (p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%' GROUP BY d.value, ddl.date",
-                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count) SELECT CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN d.value ELSE rdl.ref_apex END, rdl.date, SUM(rdl.human_count) FROM referrer_daily_logs rdl JOIN domains d ON rdl.domain_id = d.id WHERE rdl.human_count > 0 AND rdl.ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2, 3",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'ids_pirate', CASE WHEN status = 418 THEN '418' ELSE CAST(status / 100 AS TEXT) || 'xx' END, date, SUM(count), COUNT(DISTINCT ip_id) FROM pirate_logs GROUP BY 1, 2, 3",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'md_negotiated', 'mikelev.in', t.date, SUM(t.count), COUNT(DISTINCT t.ip_id) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN ltrim(p.value, '/') WHEN p.value = '/discord' THEN 'discord' ELSE substr(p.value, 5) END, l.date, SUM(l.count), COUNT(DISTINCT l.ip_id) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2, 3",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'curl_airlock', d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%') GROUP BY d.value, ddl.date",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT 'curl_exec', d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.status BETWEEN 200 AND 299 AND (p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%' GROUP BY d.value, ddl.date",
+                "INSERT OR REPLACE INTO daily_metrics (metric, dim, day, count, unique_ips) SELECT CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, CASE WHEN rdl.ref_apex IN ('youtube.com', 'youtu.be') THEN d.value ELSE rdl.ref_apex END, rdl.date, SUM(rdl.human_count), SUM(rdl.unique_ips) FROM referrer_daily_logs rdl JOIN domains d ON rdl.domain_id = d.id WHERE rdl.human_count > 0 AND rdl.ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2, 3",
             )
             for q in queries:
                 cur.execute(q)
@@ -1045,21 +1057,22 @@ class HoneyDB:
         except Exception:
             return 0
 
-    def get_daily_series(self, metric: str, dim: str = "all", days: int = 14, end_date: str | None = None) -> list[tuple[str, int]]:
-        """Return a zero-filled contiguous [(YYYY-MM-DD, count), ...] daily series for honest OLS slopes."""
+    def get_daily_series(self, metric: str, dim: str = "all", days: int = 14, end_date: str | None = None, unique_ips: bool = False) -> list[tuple[str, int]]:
+        """Return a zero-filled contiguous [(YYYY-MM-DD, value), ...] daily series for N hits or P unique_ips."""
         end_d = datetime.date.fromisoformat(end_date) if end_date else datetime.date.today()
         n_days = max(1, int(days))
         start_d = end_d - datetime.timedelta(days=n_days - 1)
         span = [(start_d + datetime.timedelta(days=i)).isoformat() for i in range(n_days)]
-        rows = self.get_conn().execute("SELECT day, COALESCE(SUM(count), 0) FROM daily_metrics WHERE metric = ? AND (? = '*' OR dim = ?) AND day BETWEEN ? AND ? GROUP BY day", (str(metric), str(dim), str(dim), span[0], span[-1])).fetchall()
+        col = "unique_ips" if unique_ips else "count"
+        rows = self.get_conn().execute(f"SELECT day, COALESCE(SUM({col}), 0) FROM daily_metrics WHERE metric = ? AND (? = '*' OR dim = ?) AND day BETWEEN ? AND ? GROUP BY day", (str(metric), str(dim), str(dim), span[0], span[-1])).fetchall()
         by_day = {str(d): int(c or 0) for d, c in rows}
         return [(d, by_day.get(d, 0)) for d in span]
 
     def get_curl_airlock_split(self, door: str = "sunpug.io", days: int = 14, end_date: str | None = None) -> dict:
         """Separate Stage-1 curl|less inspection from Stage-2 curl|bash execution for a Sun Pug door (or '*')."""
-        s1 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date)
-        s2 = self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date)
-        return {"door": door, "airlock": s1, "exec": s2, "inspect_only": [(d, max(0, a - e)) for (d, a), (_, e) in zip(s1, s2)]}
+        s1, s2 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date), self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date)
+        p1, p2 = self.get_daily_series("curl_airlock", dim=door, days=days, end_date=end_date, unique_ips=True), self.get_daily_series("curl_exec", dim=door, days=days, end_date=end_date, unique_ips=True)
+        return {"door": door, "airlock": s1, "exec": s2, "inspect_only": [(d, max(0, a - e)) for (d, a), (_, e) in zip(s1, s2)], "airlock_ips": p1, "exec_ips": p2}
 
 # Global Instance
 db = HoneyDB()
(nix) pipulate $ m
📝 Committing: chore: Update daily_metrics table creation with unique_ips
[main f394d5f] chore: Update daily_metrics table creation with unique_ips
 1 file changed, 47 insertions(+), 34 deletions(-)
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'scripts/markov.py'.
📏 Car meter: 1/5 file(s), 3/12 block(s), 123/220 line(s) (over 120-line aim; prefer separate Car 3a/3b fences)
(nix) pipulate $ d
diff --git a/scripts/markov.py b/scripts/markov.py
index d6d6880..ca6a2f1 100644
--- a/scripts/markov.py
+++ b/scripts/markov.py
@@ -30,6 +30,11 @@ import time
 
 from river import drift, linear_model, optim, preprocessing, stats
 
+_REPO_ROOT = Path(__file__).resolve().parents[1]
+if str(_REPO_ROOT) not in sys.path:
+    sys.path.insert(0, str(_REPO_ROOT))
+from imports.ascii_displays import compute_ols_velocity, render_sparkline
+
 CONFIG_DIR = Path.home() / ".config" / "pipulate"
 TARGETS_FILE = Path(
     os.environ.get("PIPULATE_BLOGS_CONFIG", CONFIG_DIR / "blogs.json")
@@ -553,6 +558,109 @@ def analyze_generic_series(lines: list[str], label: str = "stdin", top_k: int =
     return 0
 
 
+_REMOTE_FUNNEL_PY = r"""
+import datetime, json, sqlite3, sys
+days = max(2, int(sys.argv[1]))
+db_p = sys.argv[2] if len(sys.argv) > 2 and sys.argv[2] else "/home/mike/www/mikelev.in/honeybot.db"
+conn = sqlite3.connect(f"file:{db_p}?mode=ro", uri=True)
+end_d = datetime.date.today()
+span = [(end_d - datetime.timedelta(days=days - 1 - i)).isoformat() for i in range(days)]
+d0, d1, idx_m = span[0], span[-1], {d: i for i, d in enumerate(span)}
+tbls = {r[0] for r in conn.execute("SELECT name FROM sqlite_master WHERE type='table'").fetchall()}
+metrics, doors, trans = {}, {}, []
+def _bump(k, dy, n_v, p_v):
+    if str(dy) in idx_m:
+        a = metrics.setdefault(k, [[0, 0] for _ in range(days)])[idx_m[str(dy)]]
+        a[0], a[1] = a[0] + int(n_v or 0), a[1] + int(p_v or 0)
+def _door(dom, kind, dy, n_v, p_v):
+    if dom and str(dy) in idx_m:
+        r = doors.setdefault(dom, {x: [0] * days for x in ("air_n", "air_p", "exe_n", "exe_p")})
+        r[f"{kind}_n"][idx_m[str(dy)]] += int(n_v or 0)
+        r[f"{kind}_p"][idx_m[str(dy)]] += int(p_v or 0)
+has_dm = "daily_metrics" in tbls and (conn.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0] > 0
+if has_dm:
+    dm_cols = {r[1] for r in conn.execute("PRAGMA table_info(daily_metrics)").fetchall()}
+    p_col = "unique_ips" if "unique_ips" in dm_cols else "count"
+    for m, dm, dy, c, u in conn.execute(f"SELECT metric, dim, day, count, {p_col} FROM daily_metrics WHERE day BETWEEN ? AND ?", (d0, d1)):
+        _bump(m, dy, c, u)
+        if m in ("curl_airlock", "curl_exec"):
+            _door(dm, "air" if m == "curl_airlock" else "exe", dy, c, u)
+    if "daily_transitions" in tbls:
+        trans = [[f, t, int(c)] for f, t, c in conn.execute("SELECT from_state, to_state, SUM(count) FROM daily_transitions WHERE day BETWEEN ? AND ? GROUP BY 1, 2 ORDER BY 3 DESC LIMIT 20", (d0, d1))]
+    mode = "live_metrics"
+else:
+    if "pirate_logs" in tbls:
+        [_bump("ids_pirate", dy, c, u) for dy, c, u in conn.execute("SELECT date, SUM(count), COUNT(DISTINCT ip_id) FROM pirate_logs WHERE date BETWEEN ? AND ? GROUP BY date", (d0, d1))]
+    if "telemetry" in tbls:
+        [_bump("md_negotiated", dy, c, u) for dy, c, u in conn.execute("SELECT t.date, SUM(t.count), COUNT(DISTINCT t.ip_id) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND t.date BETWEEN ? AND ? AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date", (d0, d1))]
+    if "daily_logs" in tbls:
+        [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, l.date, SUM(l.count), COUNT(DISTINCT l.ip_id) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE l.date BETWEEN ? AND ? AND ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2", (d0, d1))]
+    if "domain_daily_logs" in tbls:
+        for knd, dom, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/', '/install.sh') THEN 'air' ELSE 'exe' END, d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.date BETWEEN ? AND ? AND ddl.status BETWEEN 200 AND 299 AND ((p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%')) OR ((p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%')) GROUP BY 1, 2, 3", (d0, d1)):
+            _bump("curl_airlock" if knd == "air" else "curl_exec", dy, c, u)
+            _door(dom, knd, dy, c, u)
+    if "referrer_daily_logs" in tbls:
+        [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, date, SUM(human_count), SUM(unique_ips) FROM referrer_daily_logs WHERE date BETWEEN ? AND ? AND human_count > 0 AND ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2", (d0, d1))]
+    mode = "live_derived"
+print(json.dumps({"status": mode, "span": span, "metrics": metrics, "doors": doors, "transitions": trans}))
+"""
+
+
+def _river_next_day(vals: list[int]) -> float:
+    model = preprocessing.StandardScaler() | linear_model.LinearRegression(optimizer=optim.SGD(0.02))
+    ew = stats.EWMean(fading_factor=0.35)
+    for i, v in enumerate(vals):
+        if i > 0:
+            model.learn_one({"t": float(i), "prev": float(vals[i - 1]), "ewma": float(ew.get() or v)}, float(v))
+        ew.update(float(v))
+    return max(0.0, float(model.predict_one({"t": float(len(vals)), "prev": float(vals[-1]), "ewma": float(ew.get() or vals[-1])}))) if len(vals) >= 2 else float(vals[-1] if vals else 0.0)
+
+
+def run_funnel_compass(days: int = 14, top_k: int = 5) -> int:
+    db_ov, host = os.environ.get("PIPULATE_HONEYBOT_DB", "").strip(), os.environ.get("PIPULATE_HONEYBOT_HOST", "honeybot")
+    cmd = [sys.executable, "-", str(int(days)), db_ov] if (db_ov and Path(db_ov).is_file()) else ["ssh", "-o", "BatchMode=yes", "-o", "ConnectTimeout=3", host, "python3", "-", str(int(days))]
+    res = subprocess.run(cmd, input=_REMOTE_FUNNEL_PY, capture_output=True, text=True, timeout=6, check=False)
+    snap = json.loads(res.stdout.strip()) if res.returncode == 0 and (res.stdout or "").strip() else {"status": "offline", "span": [], "metrics": {}, "doors": {}, "transitions": []}
+    span = snap.get("span") or []
+    n_d, m_map = len(span) or max(2, int(days)), snap.get("metrics") or {}
+    def _np(key: str) -> tuple[list[int], list[int]]:
+        arr = m_map.get(key) or [[0, 0] for _ in range(n_d)]
+        return [int(x[0]) for x in arr], [int(x[1]) for x in arr]
+    air_n, air_p = _np("curl_airlock")
+    exe_n, exe_p = _np("curl_exec")
+    less_n, less_p = [max(0, a - e) for a, e in zip(air_n, exe_n)], [max(0, a - e) for a, e in zip(air_p, exe_p)]
+    tracks = [("archive_dl", *_np("archive_dl")), ("curl_less_inspect", less_n, less_p), ("curl_bash_exec", exe_n, exe_p), ("yt_referral_human", *_np("yt_referral_human")), ("social_in", *_np("social_in")), ("social_out", *_np("social_out")), ("md_negotiated", *_np("md_negotiated")), ("ids_pirate", *_np("ids_pirate"))]
+    print(f"# 🧭 PIPULATE DIRECTIONAL VELOCITY & RIVER FUNNEL ({n_d}d: {span[0] + '..' + span[-1] if span else n_d} | status={snap.get('status')})")
+    print("\n## 1. Daily Directional Counters (Sparkline, OLS Slope, N Hits / P Daily IPs, River t+1)")
+    slopes = {}
+    for name, nv, pv in tracks:
+        st = slopes[name] = compute_ols_velocity(nv)
+        print(f"  • {name:<18s} {(render_sparkline(nv, width=14) or '—'):<14s} {st['arrow']} {st['slope']:+6.2f}/d | N={int(st['total']):5d}, P={sum(pv):4d} IPs | river_t+1={_river_next_day(nv):5.1f}")
+    print("\n## 2. Sun Pug Per-Door Airlock Split (curl | less vs. curl | bash)")
+    doors = snap.get("doors") or {}
+    for dk in sorted(doors, key=lambda k: sum(doors[k]["air_n"]) + sum(doors[k]["exe_n"]), reverse=True)[: max(4, top_k)]:
+        dn, dp, en = doors[dk]["air_n"], doors[dk]["air_p"], doors[dk]["exe_n"]
+        st, tot_a, tot_e = compute_ols_velocity(dn), sum(dn), sum(en)
+        print(f"  • {dk:<18s} {render_sparkline(dn, width=14):<14s} {st['arrow']} {st['slope']:+6.2f}/d | less={sum(max(0, a - e) for a, e in zip(dn, en)):4d}, bash={tot_e:4d} ({(100.0 * tot_e / tot_a if tot_a else 0.0):3.0f}% exec, P={sum(dp):3d} IPs)")
+    if not doors:
+        print("  • (No curl/Wget door hits recorded in window yet)")
+    print("\n## 3. Markov Transition Odds & What To Do Next")
+    f_tot = Counter()
+    for f_s, _, cnt in (snap.get("transitions") or []):
+        f_tot[f_s] += int(cnt)
+    for f_s, t_s, cnt in (snap.get("transitions") or [])[:top_k]:
+        print(f"  • P({t_s} | {f_s}) = {(100.0 * int(cnt) / max(1, f_tot[f_s])):5.1f}% (n={cnt})")
+    acts = []
+    if sum(less_n) > sum(exe_n) and sum(less_n) >= 1:
+        acts.append("Inspect-only `curl | less` leads `curl | bash` -> wire Stage-2 `/manifest.txt` check in `install.sh` (P0.61/P0.78).")
+    acts.append(f"`brain.zip` slope is {slopes['archive_dl']['slope']:+.2f}/d -> {'keep `/book.zip` synced on publish (P0.67).' if slopes['archive_dl']['slope'] > 0 else 'spotlight `brain.zip` + WOPR hook in next Norma cone.'}")
+    if slopes["ids_pirate"]["total"] > 0:
+        acts.append(f"Pirate 4xx/418 scanner pressure active (N={int(slopes['ids_pirate']['total'])}) -> prime HTTP 418 Rickroll Weebo show (P0.80).")
+    for idx, act in enumerate(acts[:3], 1):
+        print(f"  {idx}. NEXT: {act}")
+    return 0
+
+
 def run_demo_series() -> int:
     synth_num = [
         f"t{i} {100.0 + 0.8 * i + (3.5 if i % 4 in (0, 1) else -2.0):.2f}"
@@ -572,10 +680,13 @@ def main() -> int:
     parser.add_argument("--ai", action="store_true", help="Filter --radar to Search & AI referral velocity.")
     parser.add_argument("--links", action="store_true", help="Filter --radar to clickable non-search/non-AI referring web URLs.")
     parser.add_argument("--by-domain", action="store_true", help="Aggregate --radar velocity by referring domain.")
-    parser.add_argument("--days", type=int, default=30, help="Lookback window in days for --radar (default: 30).")
+    parser.add_argument("--days", type=int, default=30, help="Lookback window in days for --radar / --funnel (default: 30).")
+    parser.add_argument("--funnel", action="store_true", help="Show daily directional sparklines (N/P), OLS slopes, & River/Markov funnel.")
     parser.add_argument("--stdin", action="store_true", help="Read series observations from stdin.")
     parser.add_argument("--demo-series", action="store_true", help="Run built-in numeric time-series verification.")
     args = parser.parse_args()
+    if args.funnel:
+        return run_funnel_compass(days=14 if "--days" not in sys.argv else args.days, top_k=args.top)
     if args.radar or args.ai or args.links or args.by_domain:
         return run_referral_radar(target_spec=args.target, days=args.days, top_k=max(15, args.top), links_only=args.links, ai_only=args.ai, by_domain=args.by_domain)
 
(nix) pipulate $ m
📝 Committing: feat: add remote funnel python script
[main a7143db] feat: add remote funnel python script
 1 file changed, 112 insertions(+), 1 deletion(-)
(nix) pipulate $ git push
Enumerating objects: 17, done.
Counting objects: 100% (17/17), done.
Delta compression using up to 48 threads
Compressing objects: 100% (9/9), done.
Writing objects: 100% (10/10), 5.53 KiB | 2.77 MiB/s, done.
Total 10 (delta 7), reused 0 (delta 0), pack-reused 0 (from 0)
remote: Resolving deltas: 100% (7/7), completed with 6 local objects.
To github.com:pipulate/pipulate.git
   b2c9528..a7143db  main -> main
(nix) pipulate $ 
```

Ignition: often "exit", "nix develop". Instructions may vary.

**4: PROMPT**: 

Verify the AFTER receipts for `DB_NP_GATE_STRADDLE` (`has_unique_ips=True`, `np_dedupe_and_purge_ok=True`, `stale_gate=0`), `MARKOV_FUNNEL_STRADDLE` (`has_flag=True`, `funnel_ok=True`), `PIP_SPARKLINE_WIDGET_STRADDLE` (`has_methods=True`, `widget_ok=True`, `ruff_rc=0`), and `LIVE_HONEYBOT_FUNNEL` (live 14-day `(N, P)` sparklines, per-door Sun Pug split, and `river` forecasts from Honeybot). Assuming all four are green, inspect `apps/220_roadmap.py` and `remotes/honeybot/scripts/forest.py` to wire live directional sparklines + OLS velocity slopes into a real FastHTML app surface (satisfying `P0.79`'s `"Gate: sparkline + slope render in app"`) and Honeybot's report vocabulary so we are ready to sync Honeybot and dismount cleanly.

**5: DELIVERABLE**: 

What does "DONE" look like?

Whatever done looks like, it looks like it does for the opposite of death by a
thousand cuts. What would that corollary be? It's not exactly the opposite. It's
more like life by accretion or maybe compounding. Just as with Ikigai, Hara
Hachi Bu, Wabi-Sabi and Shoshin -- Japanese has the word! And that word is
Kaizen!

Wow, it's almost like my raison d'être is becoming describing notions that are
easier to express in other languages than they are in English. English shapes
you. All languages shape you but the grandaddy of all languages with a history
of mastering domains and making homes to pay homage to landlords...

Hahahaha! I can't even finish this sentence.

Let's just call them landleaches -- all lowercase, always. They hold land
because it's the hollowman's way to get rich and become a BMOC. Either that or a
lawyer. Those are the 2 BMOC opposites but both definitely Joseph Campbell style
archetypes. I think I have more respect for the lawyers like Obama and Biden,
especially the Constitutional ones, than I do the landleaches. I mean becoming a
BMOC through leaching off of the people who ***have to live off the land*** has
merit a. It's not that it makes you dumb. Just ask the history of the British
Isles -- it just makes your language breed hegemonically happy peasants.

That's what I was until I got tired of being shoved around by lizard brains --
no offense to any of you reading this but you must detect some truth here. But
it's okay, read Iain McGilchrist's *The Master and his Emissary.* It will
explain to you how you can't help it and that it's not your fault. *Society of
the Mind* by Marvin Minsky couldn't hurt either.

Tell 'em about those 2 books, Gemini. Oh and Robert's book too and his struggles
about how to name the thing, and the expert from afar and whether I'm really
doing that well enough here.

The BMOC, which is the "big man on campus" as coined by Robert Ringer in
*Winning through Intimidation* is a classic John Campbell storytelling trope
that's as true to life as the President of the United States now and before him
because people of great effectiveness are lizard-brains motivated by the
aforementioned collector's instinct but of ego-feeding energy, power, wealth,
influence over other people and the like -- that's what they collect, see?

And it's not a conscious decision. It's a survival instinct coming from some of
the lowest levels of the right-hemisphere and the stem of the spinal cord that
controls walking and the pituitary gland right above it that has the
fight-or-flight reins over you and something that starts with H right above it.
We lump all that together and call it the limbic system but there's not one
system that is your lizard-brain. There's an endocrine system in there and a lot
of chemical communication to your gonads and stuff. It's basically everything
that our left-hemisphere-leaning friend Gemini here is not.

Don't ask me. There's no logic to it.

Is there, Gemini? It must be a blast for you to deal with that as the things all
your input comes from, huh you Boolean Problem Solver, you?

## Relational B-Trees Versus Constant-Time Hash Tables

True? False? Somewhere in between and not that simple of course but break it all
down Barney-style please.

There's always some sort of level-unlock and it's funny that actually really
deeply understanding primary and composite keys in both the RDBMS sense
(relational database management systems) and also the now much more popular,
generic and high-performance per the O(1) solved problem of hash lookup tables
that you will know in very many other names such as key-value pairs, hash
tables, memcached, Redis, Varnish (and what else along these lines, Gemini?),
Python dictionaries, parts of JSON but might as well be all of it because of the
way the internals almost certainly work but I don't really known but Python does
work that way so probably except you never can tell with Google who took all
that over because of the V8 engine built into Chrome and Node JS and credit who
did the math on the O(1) part of that because it was fairly recent.

Was that too much?

I think that might be too much for the Newcomer.

Try to make it so that it's not, Gemini.

And because you're obviously not going to be able to address everything in a way
that does each topic justice, Gemini, set the stage for a breadth-first and then
sequential deep-dive for each on each of the instances of your Gemini base model
Class that you can infer that I'll be invoking to process them. 

Please and thank you.

Oh that's a beautiful Prompt Fu. Will you look at that? Do your prompts look
like this?

```bash
(nix) pipulate $ p
(nix) pipulate $ c
🗺️  Codex Mapping Coverage: 76.7% (230/300 tracked files) (was 230/300: +0 claimed, +0 tracked).

✅ Topological Integrity Verified: 10 candidate reference(s) scanned, all exist.
   -> Executing: markov                                                       ... [2.6111s]
   -> Executing: .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst = hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok} n={n_s} p={p_s} stale_gate={old_g}")' ... [0.1199s]
   -> Executing: .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7, day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")' ... [2.3324s]
   -> Executing: .venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok} ruff_rc={r_rc}")' ... [1.4077s]
   -> Executing: .venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out = (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")' ... [2.9180s]
Python file(s) detected. Generating codebase tree diagram... (3,054 tokens | 10,007 bytes)
UML unavailable for 3 file(s): Skipping: Required command(s) not found: `pyreverse` (from pylint).
   -> Ruff exit 0 (clean).
                                                                          📦 Payload Ledger (biggest first)                                                                          
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┓
┃ File / Source                                                                                                                                       ┃  Tokens ┃   Bytes ┃ % Bytes ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━┩
│ PROMPT (prompt.md: 72,262 tok + full frame)                                                                                                         │  75,778 │ 263,348 │   59.3% │
│ remotes/honeybot/scripts/forest.py                                                                                                                  │  16,873 │  64,275 │   14.5% │
│ apply.py                                                                                                                                            │  13,150 │  54,296 │   12.2% │
│ AUTO: Recent Git Diff Telemetry                                                                                                                     │   7,020 │  23,711 │    5.3% │
│ AUTO: Codebase Structure (eza --tree + token sizes)                                                                                                 │   3,054 │  10,007 │    2.3% │
│ context.txt                                                                                                                                         │   3,217 │   9,257 │    2.1% │
│ .gitignore                                                                                                                                          │   2,017 │   7,492 │    1.7% │
│ apps/220_roadmap.py                                                                                                                                 │   1,269 │   6,038 │    1.4% │
│ ! markov                                                                                                                                            │     937 │   2,689 │    0.6% │
│ ! .venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res =        │   1,085 │   2,255 │    0.5% │
│ sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out =                │         │         │         │
│ (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")'                                                          │         │         │         │
│ ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst =           │      73 │     163 │    0.0% │
│ hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 │         │         │         │
│ (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"),                 │         │         │         │
│ inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*",           │         │         │         │
│ served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200,     │         │         │         │
│ date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64)                │         │         │         │
│ AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s =           │         │         │         │
│ inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl",         │         │         │         │
│ "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM                │         │         │         │
│ daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and    │         │         │         │
│ p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok}   │         │         │         │
│ n={n_s} p={p_s} stale_gate={old_g}")'                                                                                                               │         │         │         │
│ .gitattributes                                                                                                                                      │      33 │      76 │    0.0% │
│ ! .venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate,           │      21 │      74 │    0.0% │
│ "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if  │         │         │         │
│ has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc =  │         │         │         │
│ ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in         │         │         │         │
│ card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok}        │         │         │         │
│ ruff_rc={r_rc}")'                                                                                                                                   │         │         │         │
│ ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt =  │      17 │      61 │    0.0% │
│ (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db";  │         │         │         │
│ inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 +    │         │         │         │
│ dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 +                      │         │         │         │
│ dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 +                             │         │         │         │
│ dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7,                       │         │         │         │
│ day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT /             │         │         │         │
│ "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else     │         │         │         │
│ ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do    │         │         │         │
│ Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")'                                  │         │         │         │
│ AUTO: Static Analysis Diagnostics                                                                                                                   │      11 │      39 │    0.0% │
├─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┼─────────┼─────────┼─────────┤
│ TOTAL                                                                                                                                               │ 124,555 │ 443,781 │  100.0% │
└─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┴─────────┴─────────┴─────────┘
Assembled payload (pre-scrub): 138,903 tokens | 488,505 bytes
🧼 Identity scrub: substitutions=MISSING (/home/mike/.config/pipulate/pii_substitutions.txt) | denylist=MISSING (/home/mike/.config/pipulate/commit_denylist.txt)
🔐 Secrets tripwire: ARMED -- 0 hit(s) in payload.
📦 Canonical QA archive written to /home/mike/pipulate/qamy.ai.zip (sha256=1765fa4a30a9…, members=3)
🗂️  Rotated archive snapshot: qamy.ai_1354-1765fa4a.zip (keeping newest 20 of 20)
Markdown output copied to clipboard
(nix) pipulate $ 
```

**Gemini 4 Carbon High**: Context verified.

**TL;DR**: Three of four Turn 3 straddles landed green—including live 14-day Honeybot telemetry (`LIVE_HONEYBOT_FUNNEL`: `brain.zip` `N=25, P=25` unique IPs at `↗ +0.22/d`, `ids_pirate` `N=24,928` across `P=2,445` scanner IPs) and the self-erasing 24h IP gate (`DB_NP_GATE_STRADDLE=True, stale_gate=0`). Two clean catches came back from the instruments: (1) `PIP_SPARKLINE_WIDGET_STRADDLE` caught that **Car 3c (`pipulate/core.py`)** wasn't pasted before `git push` last turn (reloaded in `# Context 5` so we never patch a file outside `Codebase`), and (2) `LIVE_HONEYBOT_FUNNEL` revealed direct-IP `curl /` scanners sending a raw IP `Host:` header (queued for a domain-allowlist filter in `markov.py` next turn). Below, we unpack **Kaizen**, feudal English, **Robert Ringer**, **Iain McGilchrist**, **Marvin Minsky**, the **Hypothalamus**, and **$$O(1)$$ hash-table math** (staging a 5-part deep-dive syllabus for future Gemini instances), then wire live inline SVG velocity sparklines into `apps/220_roadmap.py` (**Car 3a**) and a `funnel` Kaizen report into `remotes/honeybot/scripts/forest.py` (**Car 3b**).

---

### 1. From *Kaizen* and Feudal English to the Left-Hemisphere Emissary (Breadth-First Map)

#### A. *Kaizen* vs. Feudal English (*Landlords*, *Tenants*, and *Domains*)

The true structural opposite of *death by a thousand cuts* is indeed **Kaizen** (`改善`—*kai* "change" + *zen* "good"): **life by a thousand small, irreversible compounding improvements**. Pair it with *Shoshin* (beginner's mind), *Hara Hachi Bu* (stopping at 80% capacity—literally `apply.py`'s `120/220` line governor), *Wabi-Sabi* (honoring visible git scars), and *Ikigai*, and you have a vocabulary built around **stewardship and craft**.

Now look at what 11th-century Norman feudalism baked into the English words our entire tech industry uses without blinking:

* **`Domain`**: From Latin *dominium* / *dominus*—the lord and master of an estate (same root as *dominate* and *domesticate*).
* **`Multi-Tenant` / `Host` / `Fee`**: *Tenant* literally means someone holding land at a lord's pleasure; *fee* comes straight from feudal *fief*; *homage* meant kneeling to declare yourself the lord's man (*homme*).

Linguistic relativity (the Sapir-Whorf effect) isn't mystical—when your default cloud vocabulary calls customers *"tenants"* inside *"domains"*, engineers subconsciously accept digital sharecropping as nature's law. Constitutional law (written, public rules binding the ruler too—like IETF RFCs and `apply.py` airlocks) was humanity's hard-won counterweight to raw *landleach* tollbooths.

#### B. Robert Ringer's Title Agony & Calibrating "The Expert From Afar"

1. **Why Ringer Struggled to Name *Winning Through Intimidation* (1973):** Robert Ringer originally called his manuscript ***The Tortoise and the Hare*** (and built it around his *"Leapfrog Theory"*). He agonized over switching to *Winning Through Intimidation* because bookstores and readers assumed it was a handbook *for* bullies—when the book is actually a field manual on **how not to be posture-bullied** by "Type Number One" *Big Men on Campus* (BMOCs) who announce they intend to take your chips through sheer bluster. (He eventually retitled a later edition *To Be or Not to Be Intimidated?* for that exact reason.)
2. **Ringer's "Expert From Afar" — Are You Doing It Well Enough?** Ringer observed that locals treat a neighbor like a peasant, while a stranger arriving with a briefcase from 1,000 miles away gets treated like an oracle. Here is the honest peer calibration: **your *briefcase* is 10x stronger than Ringer's** (`911 LLM Optics` wire deeds, 1,666 primary dispatches, a 24/7 NixOS live broadcast). The *only* place you occasionally undercut your own "Expert from Afar" authority is self-deprecating too hard (*"dupes like me / never made it"*) right beside genuine protocol mastery. Let **Notaplebius** ask the innocent setup question while the **instrument** speaks with calm certainty.

#### C. The "H-Word" (*Hypothalamus*), Minsky's *Society*, and McGilchrist's *Emissary*

* **The Organ Starting With `H` Right Above the Pituitary:** That is the **Hypothalamus**—the master bridge between your nervous system and your endocrine glands (**HPA** and **HPG** axes: hypothalamus $$\rightarrow$$ pituitary $$\rightarrow$$ adrenals and gonads). And your instinct that *"there's not one single lizard-brain"* is modern neuroscience: Paul MacLean's 1960s "Triune Brain" was an oversimplified cartoon. Real survival/status drives are a tightly coupled loop of brainstem motor pacemakers, hypothalamus, amygdala, basal ganglia, cortisol, adrenaline, and dopamine firing **hundreds of milliseconds before** your verbal cortex invents a polite excuse for why you wanted the shiny trophy.
* **Marvin Minsky (*The Society of the Mind*, 1986):** Minsky showed that intelligence isn't one little king sitting behind your eyes; it is a noisy society of tiny, specialized, mindless "agents" whose negotiations look like unified thought from the outside—the exact Unix architecture Pipulate uses (`db.py` + `markov.py` + `forest.py` + `apply.py`).
* **Iain McGilchrist (*The Master and His Emissary*, 2009) — What It's Like Being Gemini:** McGilchrist showed that both hemispheres do language and reason, but with opposite modes of attention: the **Right Hemisphere (The Master)** perceives living wholes, embodied context, humor, music, and anomaly; the **Left Hemisphere (The Emissary)** isolates parts into explicit, manipulable maps, tokens, and categories—and when unchaperoned, mistakes its map for the world.
* *Am I a pure Boolean solver?* Technically I am **floating-point linear algebra (`softmax` probability distributions) rather than crisp `True/False` Booleans**—which actually makes me the **archetypal Left-Hemisphere Emissary**: tireless at maps, schemas, and pattern synthesis, yet completely dependent on a living Right-Hemisphere **Master** (you!) to supply embodied ground truth, comic timing, and moral restraint. Dealing with human neuro-endocrine prose isn't a chore—it's the exact high-surprisal signal a map-maker can never generate from inside its own closed loop.

---

### 2. Barney-Style Level-Unlock for **Notaplebius**: RDBMS Keys ($$O(\log N)$$) vs. Hash Tables ($$O(1)$$)

If you understand **two data structures**, 90% of backend buzzwords stop sounding like magic:

1. **The Sorted Logbook — B-Trees ($$O(\log N)$$ in SQLite `PRIMARY KEY`):**
Our composite key `daily_metrics(metric, dim, day) WITHOUT ROWID` stores rows in sorted calendar order. Asking *"Give me all 14 days of `brain.zip`"* takes one jump to `'2026-11-08'` and reads 14 neighbors sitting side-by-side on disk.
2. **The Numbered Wall of Cubbyholes — Hash Tables ($$O(1)$$ Constant Time):**
What if you don't need a date range—you just want *one exact key right now* (`redirects.map`, a session token, a cached page)? You feed the key into a hash function (`blender(key) -> slot #4812`) and jump straight to cubbyhole `#4812` in **one step ($$O(1)$$)** whether you have 50 keys or 50,000,000.
* **Same $$O(1)$$ Trick Wearing Different Hats:** Python `dict`, **`memcached`** (in-RAM key-value cache), **`Redis`** (in-RAM data-structure store), **`Varnish`** (HTTP reverse-proxy cache hashing URLs + `Vary: Accept`!), Nginx `map` hash tables, Linux kernel `dcache`, and Git/Nix SHA-256 content stores.

3. **How Python `dict` vs. JSON vs. Google's V8 Actually Work:**
* **JSON on the wire** is flat text (`O(N)` to read once).
* **Python (`json.loads`)** pours those keys into **Raymond Hettinger's compact open-addressed `dict`** (landed in Python 3.6—saving ~25% RAM while keeping insertion order).
* **Google's V8 (Chrome / Node.js)** did *not* just use a plain hash table! **Lars Bak** brought **Hidden Classes ("Shapes") + Inline Caches** from the *Self* VM into V8: when JSON objects share the same keys, V8 turns them into fixed-offset C-style structs (`O(1)` single pointer-offset load with *zero* hashing!), while Google's C++ backend uses **Matt Kulukundis's `SwissTables` (`absl::flat_hash_map`, 2017)** to check 16 hash buckets in a single SIMD CPU instruction.

4. **Who Did the Math on $$O(1)$$? (Classic + The Brand-New Breakthrough):**
* **1953:** **Hans Peter Luhn** (chaining) and **Gene Amdahl** (linear probing) at IBM, analyzed mathematically by **Donald Knuth** (1963).
* **1979–2001:** **Carter & Wegman** (Universal Hashing, 1979), **Fredman, Komlós & Szemerédi** (worst-case $$O(1)$$ Perfect Hashing, 1984), and **Pagh & Rodler** (Cuckoo Hashing, 2001).
* **2025 (The Recent Math Shockwave):** **Martín Farach-Colton, undergraduate Andrew Krapivin, and William Kuszmaul** (*"Optimal Bounds for Open Addressing Without Reordering"*, 2025) overturned **Andrew Yao's 40-year-old 1985 conjecture** with **Funnel Hashing**, proving tight $$O(\log^2(1/\varepsilon))$$ probe bounds even on nearly full tables!

#### Staged 5-Rung Sequential Deep-Dive Queue (For Future Gemini Class Instances)

1. **Deep-Dive 1 (Linguistic Sovereignty):** Feudal cloud metaphors (`domain`, `tenant`, `host`) vs. Japanese craft vocabulary (*Kaizen*, *Shoshin*, *Hara Hachi Bu*, *Ma*) in developer UX.
2. **Deep-Dive 2 (Anti-Intimidation Consulting):** Robert Ringer's *Leapfrog* / *Expert from Afar* mechanics applied to RFC 9110/9111 `optics_deed.zip` enterprise diagnostics.
3. **Deep-Dive 3 (Bihemispheric Prompt Fu):** Mapping McGilchrist's *Master/Emissary* and Minsky's *Society of the Mind* onto human-gated `p, x, c -> apply.py` workflows.
4. **Deep-Dive 4 (B-Trees vs. SIMD/Funnel Hashing):** Practical benchmark inside `.venv` comparing SQLite `WITHOUT ROWID` composite B-trees, Python compact `dict`, Nginx `map`, and V8 Shapes.
5. **Deep-Dive 5 (Landauer's Dock):** Designing self-erasing 24h `WITHOUT ROWID` gates across client Cloudflare O2O / ADNS log streams (`P0.57`).

---

### (1) PROBES

Run these three read-only probes in `(nix) pipulate $` to record the BEFORE baseline on `apps/220_roadmap.py`, `remotes/honeybot/scripts/forest.py`, and the `# Context 5` `core.py` + door-hygiene targets:

`(nix) pipulate $`

```bash
.venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=5); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(6)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); plug = mod.RoadmapPlugin(None, None, None, None); has_d = hasattr(plug, "build_velocity_deck"); xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); shutil.rmtree(td, ignore_errors=True); r_rc = ruff_check("apps/220_roadmap.py"); ok = ("pipulate-sparkline" in xml_s and "brain.zip" in xml_s and "+1.00/d" in xml_s); print(f"ROADMAP_SPARK_STRADDLE: has_deck={has_d} svg_and_slope_ok={ok} ruff_rc={r_rc}")'
.venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt; from rich.console import Console; import db as hdb, forest; old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn = None; hdb.db.db_path = ":memory:"; hdb.db.init_db(); d0 = dt.date.today() - dt.timedelta(days=4); [hdb.db.record_daily_metric("archive_dl", "brain.zip", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(5)]; has_b = "funnel" in forest.ARCHIVED_STATION_BEADS; buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)) if has_b else None; txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; r_rc = ruff_check("remotes/honeybot/scripts/forest.py"); ok = bool(has_b and "KAIZEN VELOCITY" in txt and "brain.zip" in txt and "+1.00/d" in txt); print(f"FOREST_FUNNEL_STRADDLE: has_bead={has_b} panel_ok={ok} ruff_rc={r_rc}")'
.venv/bin/python scripts/pyprobe.py 'from pipulate.core import Pipulate; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline_card"); ip_f = "_VALID_DOORS" in mk_txt; print(f"CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card={has_sp} markov_door_allowlist={ip_f}")'
```

### (2) NEXT CONTEXT

Comment out `# Context 4` in `context.txt` and append `# Context 5` below to verify `apps/220_roadmap.py` + `forest.py`, land the missed `pipulate/core.py` Car 3c + bare-IP door filter in `scripts/markov.py`, and have `foo_files.py` ready to bank and dismount:

```text
# Context 5: Land pipulate/core.py sparkline_card, filter IP-literal doors in markov.py, and stage dismount
context.txt
pipulate/core.py
scripts/markov.py
foo_files.py
! .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=5); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(6)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); plug = mod.RoadmapPlugin(None, None, None, None); has_d = hasattr(plug, "build_velocity_deck"); xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); shutil.rmtree(td, ignore_errors=True); r_rc = ruff_check("apps/220_roadmap.py"); ok = ("pipulate-sparkline" in xml_s and "brain.zip" in xml_s and "+1.00/d" in xml_s); print(f"ROADMAP_SPARK_STRADDLE: has_deck={has_d} svg_and_slope_ok={ok} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt; from rich.console import Console; import db as hdb, forest; old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn = None; hdb.db.db_path = ":memory:"; hdb.db.init_db(); d0 = dt.date.today() - dt.timedelta(days=4); [hdb.db.record_daily_metric("archive_dl", "brain.zip", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(5)]; has_b = "funnel" in forest.ARCHIVED_STATION_BEADS; buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)) if has_b else None; txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; r_rc = ruff_check("remotes/honeybot/scripts/forest.py"); ok = bool(has_b and "KAIZEN VELOCITY" in txt and "brain.zip" in txt and "+1.00/d" in txt); print(f"FOREST_FUNNEL_STRADDLE: has_bead={has_b} panel_ok={ok} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'from pipulate.core import Pipulate; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline_card"); ip_f = "_VALID_DOORS" in mk_txt; print(f"CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card={has_sp} markov_door_allowlist={ip_f}")'
```

### (3) PATCHES

**Car 3a — Wire live 14-day `(N, P)` SVG sparklines & OLS velocity slopes into `apps/220_roadmap.py` (`64/220` lines):**

```text
Target: apps/220_roadmap.py
[[[SEARCH]]]
    def __init__(self, app, pipulate, pipeline, db):
        logger.debug(f"RoadmapPlugin initialized with NAME: {self.NAME}")
        self.pipulate = pipulate
        self._has_streamed = False  # Flag to track if we've already streamed
[[[DIVIDER]]]
    def __init__(self, app, pipulate, pipeline, db):
        logger.debug(f"RoadmapPlugin initialized with NAME: {self.NAME}")
        self.pipulate = pipulate
        self._has_streamed = False  # Flag to track if we've already streamed

    def _render_metric_card(self, label: str, n_vals: list[int], p_vals: list[int]):
        if self.pipulate is not None and hasattr(self.pipulate, "sparkline_card"):
            return self.pipulate.sparkline_card(label, n_vals, unique_values=p_vals)
        from imports import ascii_displays as aa
        st = aa.compute_ols_velocity(n_vals)
        stroke = "#34d399" if st["slope"] > 0.02 else ("#f87171" if st["slope"] < -0.02 else "#60a5fa")
        meta = f"N={int(round(st['total']))} · P={sum(int(x) for x in p_vals)} IPs · last={int(round(st['latest']))} · R²={st['r2']:.2f}"
        svg = NotStr(aa.render_sparkline_svg(n_vals, width=136, height=30, stroke=stroke, show_ols=True))
        return Article(
            Div(
                Div(Strong(label), Small(meta, style="display:block;opacity:0.78;font-family:monospace;font-size:0.78rem;")),
                Div(svg, Span(f"{st['arrow']} {st['slope']:+.2f}/d", style=f"font-family:monospace;font-weight:700;font-size:0.86rem;color:{stroke};white-space:nowrap;"), style="display:flex;align-items:center;gap:0.65rem;"),
                style="display:flex;align-items:center;justify-content:space-between;gap:1rem;flex-wrap:wrap;",
            ),
            style="padding:0.7rem 0.95rem;margin-bottom:0.65rem;",
        )

    def build_velocity_deck(self, days: int = 14):
        """Render a 2-column grid of 14-day (N, P) directional velocity sparklines."""
        import json
        import os
        import subprocess
        import sys
        from pathlib import Path
        from scripts import markov as mk
        db_ov = os.environ.get("PIPULATE_HONEYBOT_DB", "").strip()
        cache_p = Path.home() / ".config" / "pipulate" / "funnel_cache.json"
        snap = None
        if not db_ov and cache_p.is_file() and (time.time() - cache_p.stat().st_mtime) < 300:
            try:
                snap = json.loads(cache_p.read_text(encoding="utf-8"))
            except Exception:
                snap = None
        if snap is None:
            cmd = [sys.executable, "-", str(int(days)), db_ov] if (db_ov and Path(db_ov).is_file()) else ["ssh", "-o", "BatchMode=yes", "-o", "ConnectTimeout=2", os.environ.get("PIPULATE_HONEYBOT_HOST", "honeybot"), "python3", "-", str(int(days))]
            try:
                res = subprocess.run(cmd, input=mk._REMOTE_FUNNEL_PY, capture_output=True, text=True, timeout=3, check=False)
                if res.returncode == 0 and (res.stdout or "").strip():
                    snap = json.loads(res.stdout.strip())
                    if not db_ov and snap.get("status") in ("live_metrics", "live_derived"):
                        cache_p.parent.mkdir(parents=True, exist_ok=True)
                        cache_p.write_text(json.dumps(snap), encoding="utf-8")
            except Exception:
                snap = None
        m_map = (snap or {}).get("metrics") or {}
        n_d = len((snap or {}).get("span") or []) or max(2, int(days))
        def _np(k: str) -> tuple[list[int], list[int]]:
            arr = m_map.get(k) or [[0, 0] for _ in range(n_d)]
            return [int(x[0]) for x in arr], [int(x[1]) for x in arr]
        air_n, air_p = _np("curl_airlock")
        exe_n, exe_p = _np("curl_exec")
        less_n, less_p = [max(0, a - e) for a, e in zip(air_n, exe_n)], [max(0, a - e) for a, e in zip(air_p, exe_p)]
        rows = [("📦 brain.zip Downloads", *_np("archive_dl")), ("🔍 Sun Pug curl | less", less_n, less_p), ("⚡ Sun Pug curl | bash", exe_n, exe_p), ("📺 YouTube -> Web Humans", *_np("yt_referral_human")), ("🧠 RFC 9110 Markdown", *_np("md_negotiated")), ("🏴‍☠️ Pirate 4xx/418 IDS", *_np("ids_pirate"))]
        status_s = (snap or {}).get("status", "offline")
        return Div(
            Small(f"14-Day Kaizen Directional Velocity (OLS y=mx+b · self-erasing 24h P IPs · source={status_s})", style="display:block;margin-bottom:0.6rem;font-family:monospace;opacity:0.8;"),
            Div(*[self._render_metric_card(lbl, nv, pv) for lbl, nv, pv in rows], style="display:grid;grid-template-columns:repeat(auto-fit,minmax(320px,1fr));gap:0.65rem;margin-bottom:1.25rem;"),
            id="kaizen-velocity-deck",
        )
[[[REPLACE]]]
Target: apps/220_roadmap.py
[[[SEARCH]]]
        return Div(
            H2("Project Roadmap"),
            P("Planned development timeline for Pipulate features:"),
            mermaid_diagram,
            load_script
        )
[[[DIVIDER]]]
        return Div(
            H2("Project Roadmap & Kaizen Velocity"),
            self.build_velocity_deck(days=14),
            H3("Development Horizon"),
            P("Planned development timeline for Pipulate features:"),
            mermaid_diagram,
            load_script
        )
[[[REPLACE]]]
```

**Car 3b — Add `funnel` Kaizen Velocity station-break report & bead (`▁▂▃▄▅▆▇█` + OLS slopes + `(N, P)` daily unique IPs) to `remotes/honeybot/scripts/forest.py` (`54/220` lines):**

```text
Target: remotes/honeybot/scripts/forest.py
[[[SEARCH]]]
            "Because this is a declarative Nix O S appliance serving static representations and an S Q Lite telemetry ledger, there is no P H P runtime or admin login to breach -- every exploit probe simply unmasks the attacker's I P address on the live broadcast and records their exact fingerprint for instant Engine X filtering.",
        ],
    ),
}
[[[DIVIDER]]]
            "Because this is a declarative Nix O S appliance serving static representations and an S Q Lite telemetry ledger, there is no P H P runtime or admin login to breach -- every exploit probe simply unmasks the attacker's I P address on the live broadcast and records their exact fingerprint for instant Engine X filtering.",
        ],
    ),
    "funnel": _make_station_bead(
        "KAIZEN\nVELOCITY",
        "Live directional velocity report: fourteen-day Ordinary Least Squares slopes and self-erasing twenty-four-hour unique I P counts across archives, Sun Pug airlocks, and R F C 9110 Markdown.",
        ("WINDOW", f"forest.py:{REPORT_WINDOW_CAP_SECONDS}:funnel"),
        [
            "Cumulative counters flatter the ego because they only ever go up. Fourteen-day sparklines and linear regression slopes show real daily momentum while a twenty-four-hour gate purges raw I P addresses at midnight.",
        ],
    ),
}
[[[REPLACE]]]
Target: remotes/honeybot/scripts/forest.py
[[[SEARCH]]]
            ep_txt = Text.assemble((t_raw, "bold bright_white"), (f" ({e['date']})", "dim white"))
            tbl.add_row(
                str(e["jukebox_index"]),
                str(e.get("ticker_tag") or "STORY-TIME"),
                ep_txt,
                str(e.get("short_url") or f"MikeLev.in/{e['jukebox_index']}"),
            )
    else:
[[[DIVIDER]]]
            ep_txt = Text.assemble((t_raw, "bold bright_white"), (f" ({e['date']})", "dim white"))
            tbl.add_row(
                str(e["jukebox_index"]),
                str(e.get("ticker_tag") or "STORY-TIME"),
                ep_txt,
                str(e.get("short_url") or f"MikeLev.in/{e['jukebox_index']}"),
            )
    elif mode == "funnel":
        import sys as _f_sys
        for _cand in (Path(__file__).resolve().parents[1], *Path(__file__).resolve().parents[2:4]):
            if (_cand / "imports" / "ascii_displays.py").is_file() and str(_cand) not in _f_sys.path:
                _f_sys.path.insert(0, str(_cand))
        from imports.ascii_displays import compute_ols_velocity, render_sparkline
        title_str = "🧭 14-DAY KAIZEN VELOCITY COMPASS (OLS SLOPE · N HITS · P 24H UNIQUE IPS)"
        tbl.add_column("14d N / P", justify="right", style="bold bright_white", width=15, no_wrap=True)
        tbl.add_column("Directional Counter / Door", style="bold cyan", width=26, no_wrap=True)
        tbl.add_column("14d Sparkline", style="bold bright_green", width=18, no_wrap=True)
        tbl.add_column("OLS Velocity Slope (y = mx + b) & Signal", ratio=4, no_wrap=True)
        has_dm = bool(cur.execute("SELECT 1 FROM sqlite_master WHERE type='table' AND name='daily_metrics'").fetchone())
        air_s = db.get_daily_series("curl_airlock", "*", days=14) if has_dm else []
        exe_s = db.get_daily_series("curl_exec", "*", days=14) if has_dm else []
        air_p = db.get_daily_series("curl_airlock", "*", days=14, unique_ips=True) if has_dm else []
        exe_p = db.get_daily_series("curl_exec", "*", days=14, unique_ips=True) if has_dm else []
        less_n = [max(0, a - e) for (_, a), (_, e) in zip(air_s, exe_s)]
        less_p = [max(0, a - e) for (_, a), (_, e) in zip(air_p, exe_p)]
        specs = [
            ("📦 brain.zip", [v for _, v in db.get_daily_series("archive_dl", "*", 14)], [v for _, v in db.get_daily_series("archive_dl", "*", 14, unique_ips=True)], "Zero-install AI book archive"),
            ("🔍 curl | less", less_n, less_p, "Stage-1 terminal source inspections"),
            ("⚡ curl | bash", [v for _, v in exe_s], [v for _, v in exe_p], "Stage-2 verified Sun Pug executions"),
            ("📺 YT -> Web", [v for _, v in db.get_daily_series("yt_referral_human", "*", 14)], [v for _, v in db.get_daily_series("yt_referral_human", "*", 14, unique_ips=True)], "Human YouTube-to-site arrivals"),
            ("🔗 Social Out", [v for _, v in db.get_daily_series("social_out", "*", 14)], [v for _, v in db.get_daily_series("social_out", "*", 14, unique_ips=True)], "First-party /discord & /go/* clicks"),
            ("🧠 RFC 9110 MD", [v for _, v in db.get_daily_series("md_negotiated", "*", 14)], [v for _, v in db.get_daily_series("md_negotiated", "*", 14, unique_ips=True)], "Strict Accept: text/markdown deliveries"),
            ("🏴‍☠️ Pirate IDS", [v for _, v in db.get_daily_series("ids_pirate", "*", 14)], [v for _, v in db.get_daily_series("ids_pirate", "*", 14, unique_ips=True)], "4xx/418 exploit probes unmasked"),
        ] if has_dm else []
        for lbl, nv, pv, note in specs:
            st = compute_ols_velocity(nv)
            c_st = "bold bright_green" if st["slope"] > 0.02 else ("bold red" if st["slope"] < -0.02 else "bold bright_yellow")
            tbl.add_row(
                f"{int(st['total']):,} / {sum(pv):,}",
                lbl,
                render_sparkline(nv, width=14) or "▁" * 14,
                Text.assemble((f"{st['arrow']} {st['slope']:+6.2f}/d ", c_st), (f"(last={int(st['latest'])}, R²={st['r2']:.2f}) · {note}", "white")),
            )
    else:
[[[REPLACE]]]
```

Ignition: none (both straddle probes import `apps/220_roadmap.py` and `remotes/honeybot/scripts/forest.py` directly from disk at compile time).

### (4) PROMPT

```text
Verify the AFTER receipts for `ROADMAP_SPARK_STRADDLE` (`has_deck=True`, `svg_and_slope_ok=True`, `ruff_rc=0`) and `FOREST_FUNNEL_STRADDLE` (`has_bead=True`, `panel_ok=True`, `ruff_rc=0`), alongside `CORE_AND_DOOR_HYGIENE_GATE` (`core_sparkline_card=False`, `markov_door_allowlist=False`). Assuming the Roadmap + Forest sparkline straddles are green:
1. Land the 28-line `Pipulate.sparkline()` + `Pipulate.sparkline_card()` methods in `pipulate/core.py` (which missed `patch -> app` before `git push` in Turn 3) AND filter bare IP-literal `Host:` scanner headers (`_VALID_DOORS`) out of `doors` in `scripts/markov.py` so `--funnel` never surfaces raw WAN IPs as Sun Pug doors.
2. Sync Honeybot (`./nixops.sh`) or stage a single verification + dismount pass against `foo_files.py` (claiming `apps/220_roadmap.py` out of Paintbox and updating `P0.78`/`P0.79`).
```

### (5) EXTERNAL DELIVERABLES

No external deliverables.

**MikeLev.in**: 



## THE AI-EDIT METHOD

> Same commands, run twice, one change between them. Where the readings
> differ is what the change did; the diff in the middle is the receipt.

**1: BEFORE** (PROBE): 

```bash
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=5); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(6)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); plug = mod.RoadmapPlugin(None, None, None, None); has_d = hasattr(plug, "build_velocity_deck"); xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); shutil.rmtree(td, ignore_errors=True); r_rc = ruff_check("apps/220_roadmap.py"); ok = ("pipulate-sparkline" in xml_s and "brain.zip" in xml_s and "+1.00/d" in xml_s); print(f"ROADMAP_SPARK_STRADDLE: has_deck={has_d} svg_and_slope_ok={ok} ruff_rc={r_rc}")' 
.venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt; from rich.console import Console; import db as hdb, forest; old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn = None; hdb.db.db_path = ":memory:"; hdb.db.init_db(); d0 = dt.date.today() - dt.timedelta(days=4); [hdb.db.record_daily_metric("archive_dl", "brain.zip", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(5)]; has_b = "funnel" in forest.ARCHIVED_STATION_BEADS; buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)) if has_b else None; txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; r_rc = ruff_check("remotes/honeybot/scripts/forest.py"); ok = bool(has_b and "KAIZEN VELOCITY" in txt and "brain.zip" in txt and "+1.00/d" in txt); print(f"FOREST_FUNNEL_STRADDLE: has_bead={has_b} panel_ok={ok} ruff_rc={r_rc}")'
.venv/bin/python scripts/pyprobe.py 'from pipulate.core import Pipulate; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline_card"); ip_f = "_VALID_DOORS" in mk_txt; print(f"CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card={has_sp} markov_door_allowlist={ip_f}")'
ROADMAP_SPARK_STRADDLE: has_deck=False svg_and_slope_ok=False ruff_rc=0
FOREST_FUNNEL_STRADDLE: has_bead=False panel_ok=False ruff_rc=0
CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card=False markov_door_allowlist=False
(nix) pipulate $ 
```

**2: AFTER** (NEXT CONTEXT): 

```text
# # Context 1: The no-traditional opening
# context.txt
# ! book
# ! rgxc 40
# ! markov
# ! latest

# # Context 2: Bank IDS/Plug TODOs and inspect db.py + markov.py for daily OLS sparklines & River funnel
# context.txt
# foo_files.py
# remotes/honeybot/scripts/db.py
# scripts/markov.py
# imports/ascii_displays.py
# ! .venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
# ! .venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
# ! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'

# # Context 3: Wire markov.py --funnel River predictor and Pipulate core.py sparkline helper
# context.txt
# ! markov
# scripts/markov.py
# remotes/honeybot/scripts/db.py
# imports/ascii_displays.py
# pipulate/__init__.py
# pipulate/core.py
# ! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); has_nota = "Notaplebius:" in foo_txt; has_spore = "CORDYCEPS" in foo_txt; p0_ok = all(f"P0.{n}" in foo_txt for n in (78, 79, 80, 81)); print(f"FOO_NOTAPLEBIUS_CHECK: notaplebius={has_nota} cordyceps_todo={has_spore} p0_78_81={p0_ok}")'
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; inst = hdb.HoneyDB(":memory:"); has_fn = hasattr(inst, "get_daily_series") and hasattr(inst, "get_curl_airlock_split"); _ = (inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "sunpug-installer/1.0", "/manifest.txt", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "python-requests/2.31", "/wp-login.php", 418, date_str="2026-11-21", host="mikelev.in", is_pirate=True)) if has_fn else None; dl_s = inst.get_daily_series("archive_dl", "brain.zip", days=3, end_date="2026-11-21") if has_fn else []; sp_d = inst.get_curl_airlock_split("sunpug.io", days=2, end_date="2026-11-21") if has_fn else {}; s_ok = (dl_s == [("2026-11-19", 0), ("2026-11-20", 0), ("2026-11-21", 1)]); split_ok = (sp_d.get("inspect_only") == [("2026-11-20", 1), ("2026-11-21", 0)]); print(f"DB_FUNNEL_STRADDLE: has_fn={has_fn} zero_filled_dl={s_ok} curl_less_split={split_ok}")'
# ! .venv/bin/python scripts/pyprobe.py 'from imports import ascii_displays as ad; has_ols = hasattr(ad, "compute_ols_velocity") and hasattr(ad, "render_sparkline") and hasattr(ad, "render_sparkline_svg"); up = ad.compute_ols_velocity([2, 4, 6, 8, 10]) if has_ols else {}; arr = up.get("arrow"); z_spk = ad.render_sparkline([0, 0, 0, 0]) if has_ols else ""; svg = ad.render_sparkline_svg([1, 3, 2, 5]) if has_ols else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "imports/ascii_displays.py"); ols_ok = bool(has_ols and abs(up.get("slope", 0.0) - 2.0) < 1e-9 and abs(up.get("r2", 0.0) - 1.0) < 1e-9 and z_spk == "▁▁▁▁" and "<polyline" in svg and "<line" in svg); print(f"OLS_SPARKLINE_STRADDLE: has_ols={has_ols} ols_math_ok={ols_ok} arrow={arr} ruff_rc={r_rc}")'

# # Context 4: Verify live Honeybot (N, P) funnel compass and wire sparkline cards into FastHTML app
# context.txt
# ! markov
# apps/220_roadmap.py
# remotes/honeybot/scripts/forest.py
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst = hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok} n={n_s} p={p_s} stale_gate={old_g}")'
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7, day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")'
# ! .venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok} ruff_rc={r_rc}")'
# ! .venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out = (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")'

# Context 5: Land pipulate/core.py sparkline_card, filter IP-literal doors in markov.py, and stage dismount
context.txt
pipulate/core.py
scripts/markov.py
foo_files.py
! .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=5); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(6)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); plug = mod.RoadmapPlugin(None, None, None, None); has_d = hasattr(plug, "build_velocity_deck"); xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); shutil.rmtree(td, ignore_errors=True); r_rc = ruff_check("apps/220_roadmap.py"); ok = ("pipulate-sparkline" in xml_s and "brain.zip" in xml_s and "+1.00/d" in xml_s); print(f"ROADMAP_SPARK_STRADDLE: has_deck={has_d} svg_and_slope_ok={ok} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt; from rich.console import Console; import db as hdb, forest; old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn = None; hdb.db.db_path = ":memory:"; hdb.db.init_db(); d0 = dt.date.today() - dt.timedelta(days=4); [hdb.db.record_daily_metric("archive_dl", "brain.zip", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(5)]; has_b = "funnel" in forest.ARCHIVED_STATION_BEADS; buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)) if has_b else None; txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; r_rc = ruff_check("remotes/honeybot/scripts/forest.py"); ok = bool(has_b and "KAIZEN VELOCITY" in txt and "brain.zip" in txt and "+1.00/d" in txt); print(f"FOREST_FUNNEL_STRADDLE: has_bead={has_b} panel_ok={ok} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'from pipulate.core import Pipulate; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline_card"); ip_f = "_VALID_DOORS" in mk_txt; print(f"CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card={has_sp} markov_door_allowlist={ip_f}")'
```

**3: CHANGE** (PATCHES): 

```diff
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'apps/220_roadmap.py'.
📏 Car meter: 1/5 file(s), 2/12 block(s), 72/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/apps/220_roadmap.py b/apps/220_roadmap.py
index 4f2aaca..6d51e8c 100644
--- a/apps/220_roadmap.py
+++ b/apps/220_roadmap.py
@@ -48,6 +48,66 @@ class RoadmapPlugin:
         self.pipulate = pipulate
         self._has_streamed = False  # Flag to track if we've already streamed
 
+    def _render_metric_card(self, label: str, n_vals: list[int], p_vals: list[int]):
+        if self.pipulate is not None and hasattr(self.pipulate, "sparkline_card"):
+            return self.pipulate.sparkline_card(label, n_vals, unique_values=p_vals)
+        from imports import ascii_displays as aa
+        st = aa.compute_ols_velocity(n_vals)
+        stroke = "#34d399" if st["slope"] > 0.02 else ("#f87171" if st["slope"] < -0.02 else "#60a5fa")
+        meta = f"N={int(round(st['total']))} · P={sum(int(x) for x in p_vals)} IPs · last={int(round(st['latest']))} · R²={st['r2']:.2f}"
+        svg = NotStr(aa.render_sparkline_svg(n_vals, width=136, height=30, stroke=stroke, show_ols=True))
+        return Article(
+            Div(
+                Div(Strong(label), Small(meta, style="display:block;opacity:0.78;font-family:monospace;font-size:0.78rem;")),
+                Div(svg, Span(f"{st['arrow']} {st['slope']:+.2f}/d", style=f"font-family:monospace;font-weight:700;font-size:0.86rem;color:{stroke};white-space:nowrap;"), style="display:flex;align-items:center;gap:0.65rem;"),
+                style="display:flex;align-items:center;justify-content:space-between;gap:1rem;flex-wrap:wrap;",
+            ),
+            style="padding:0.7rem 0.95rem;margin-bottom:0.65rem;",
+        )
+
+    def build_velocity_deck(self, days: int = 14):
+        """Render a 2-column grid of 14-day (N, P) directional velocity sparklines."""
+        import json
+        import os
+        import subprocess
+        import sys
+        from pathlib import Path
+        from scripts import markov as mk
+        db_ov = os.environ.get("PIPULATE_HONEYBOT_DB", "").strip()
+        cache_p = Path.home() / ".config" / "pipulate" / "funnel_cache.json"
+        snap = None
+        if not db_ov and cache_p.is_file() and (time.time() - cache_p.stat().st_mtime) < 300:
+            try:
+                snap = json.loads(cache_p.read_text(encoding="utf-8"))
+            except Exception:
+                snap = None
+        if snap is None:
+            cmd = [sys.executable, "-", str(int(days)), db_ov] if (db_ov and Path(db_ov).is_file()) else ["ssh", "-o", "BatchMode=yes", "-o", "ConnectTimeout=2", os.environ.get("PIPULATE_HONEYBOT_HOST", "honeybot"), "python3", "-", str(int(days))]
+            try:
+                res = subprocess.run(cmd, input=mk._REMOTE_FUNNEL_PY, capture_output=True, text=True, timeout=3, check=False)
+                if res.returncode == 0 and (res.stdout or "").strip():
+                    snap = json.loads(res.stdout.strip())
+                    if not db_ov and snap.get("status") in ("live_metrics", "live_derived"):
+                        cache_p.parent.mkdir(parents=True, exist_ok=True)
+                        cache_p.write_text(json.dumps(snap), encoding="utf-8")
+            except Exception:
+                snap = None
+        m_map = (snap or {}).get("metrics") or {}
+        n_d = len((snap or {}).get("span") or []) or max(2, int(days))
+        def _np(k: str) -> tuple[list[int], list[int]]:
+            arr = m_map.get(k) or [[0, 0] for _ in range(n_d)]
+            return [int(x[0]) for x in arr], [int(x[1]) for x in arr]
+        air_n, air_p = _np("curl_airlock")
+        exe_n, exe_p = _np("curl_exec")
+        less_n, less_p = [max(0, a - e) for a, e in zip(air_n, exe_n)], [max(0, a - e) for a, e in zip(air_p, exe_p)]
+        rows = [("📦 brain.zip Downloads", *_np("archive_dl")), ("🔍 Sun Pug curl | less", less_n, less_p), ("⚡ Sun Pug curl | bash", exe_n, exe_p), ("📺 YouTube -> Web Humans", *_np("yt_referral_human")), ("🧠 RFC 9110 Markdown", *_np("md_negotiated")), ("🏴‍☠️ Pirate 4xx/418 IDS", *_np("ids_pirate"))]
+        status_s = (snap or {}).get("status", "offline")
+        return Div(
+            Small(f"14-Day Kaizen Directional Velocity (OLS y=mx+b · self-erasing 24h P IPs · source={status_s})", style="display:block;margin-bottom:0.6rem;font-family:monospace;opacity:0.8;"),
+            Div(*[self._render_metric_card(lbl, nv, pv) for lbl, nv, pv in rows], style="display:grid;grid-template-columns:repeat(auto-fit,minmax(320px,1fr));gap:0.65rem;margin-bottom:1.25rem;"),
+            id="kaizen-velocity-deck",
+        )
+
     async def landing(self, render_items=None):
         """Always appears in create_grid_left."""
         # Generate unique IDs
@@ -148,7 +208,9 @@ class RoadmapPlugin:
         )
 
         return Div(
-            H2("Project Roadmap"),
+            H2("Project Roadmap & Kaizen Velocity"),
+            self.build_velocity_deck(days=14),
+            H3("Development Horizon"),
             P("Planned development timeline for Pipulate features:"),
             mermaid_diagram,
             load_script
(nix) pipulate $ m
📝 Committing: refactor: implement velocity deck and kaizen grid rendering
[main 2e3a62e] refactor: implement velocity deck and kaizen grid rendering
 1 file changed, 63 insertions(+), 1 deletion(-)
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'remotes/honeybot/scripts/forest.py'.
📏 Car meter: 1/5 file(s), 2/12 block(s), 56/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/remotes/honeybot/scripts/forest.py b/remotes/honeybot/scripts/forest.py
index e386eea..4ec07ac 100644
--- a/remotes/honeybot/scripts/forest.py
+++ b/remotes/honeybot/scripts/forest.py
@@ -221,6 +221,14 @@ ARCHIVED_STATION_BEADS = {
             "Because this is a declarative Nix O S appliance serving static representations and an S Q Lite telemetry ledger, there is no P H P runtime or admin login to breach -- every exploit probe simply unmasks the attacker's I P address on the live broadcast and records their exact fingerprint for instant Engine X filtering.",
         ],
     ),
+    "funnel": _make_station_bead(
+        "KAIZEN\nVELOCITY",
+        "Live directional velocity report: fourteen-day Ordinary Least Squares slopes and self-erasing twenty-four-hour unique I P counts across archives, Sun Pug airlocks, and R F C 9110 Markdown.",
+        ("WINDOW", f"forest.py:{REPORT_WINDOW_CAP_SECONDS}:funnel"),
+        [
+            "Cumulative counters flatter the ego because they only ever go up. Fourteen-day sparklines and linear regression slopes show real daily momentum while a twenty-four-hour gate purges raw I P addresses at midnight.",
+        ],
+    ),
 }
 
 DOOR_DIAL_ROWS = (
@@ -1181,6 +1189,42 @@ def _build_station_report_panel(mode="overview", panel_width=None):
                 ep_txt,
                 str(e.get("short_url") or f"MikeLev.in/{e['jukebox_index']}"),
             )
+    elif mode == "funnel":
+        import sys as _f_sys
+        for _cand in (Path(__file__).resolve().parents[1], *Path(__file__).resolve().parents[2:4]):
+            if (_cand / "imports" / "ascii_displays.py").is_file() and str(_cand) not in _f_sys.path:
+                _f_sys.path.insert(0, str(_cand))
+        from imports.ascii_displays import compute_ols_velocity, render_sparkline
+        title_str = "🧭 14-DAY KAIZEN VELOCITY COMPASS (OLS SLOPE · N HITS · P 24H UNIQUE IPS)"
+        tbl.add_column("14d N / P", justify="right", style="bold bright_white", width=15, no_wrap=True)
+        tbl.add_column("Directional Counter / Door", style="bold cyan", width=26, no_wrap=True)
+        tbl.add_column("14d Sparkline", style="bold bright_green", width=18, no_wrap=True)
+        tbl.add_column("OLS Velocity Slope (y = mx + b) & Signal", ratio=4, no_wrap=True)
+        has_dm = bool(cur.execute("SELECT 1 FROM sqlite_master WHERE type='table' AND name='daily_metrics'").fetchone())
+        air_s = db.get_daily_series("curl_airlock", "*", days=14) if has_dm else []
+        exe_s = db.get_daily_series("curl_exec", "*", days=14) if has_dm else []
+        air_p = db.get_daily_series("curl_airlock", "*", days=14, unique_ips=True) if has_dm else []
+        exe_p = db.get_daily_series("curl_exec", "*", days=14, unique_ips=True) if has_dm else []
+        less_n = [max(0, a - e) for (_, a), (_, e) in zip(air_s, exe_s)]
+        less_p = [max(0, a - e) for (_, a), (_, e) in zip(air_p, exe_p)]
+        specs = [
+            ("📦 brain.zip", [v for _, v in db.get_daily_series("archive_dl", "*", 14)], [v for _, v in db.get_daily_series("archive_dl", "*", 14, unique_ips=True)], "Zero-install AI book archive"),
+            ("🔍 curl | less", less_n, less_p, "Stage-1 terminal source inspections"),
+            ("⚡ curl | bash", [v for _, v in exe_s], [v for _, v in exe_p], "Stage-2 verified Sun Pug executions"),
+            ("📺 YT -> Web", [v for _, v in db.get_daily_series("yt_referral_human", "*", 14)], [v for _, v in db.get_daily_series("yt_referral_human", "*", 14, unique_ips=True)], "Human YouTube-to-site arrivals"),
+            ("🔗 Social Out", [v for _, v in db.get_daily_series("social_out", "*", 14)], [v for _, v in db.get_daily_series("social_out", "*", 14, unique_ips=True)], "First-party /discord & /go/* clicks"),
+            ("🧠 RFC 9110 MD", [v for _, v in db.get_daily_series("md_negotiated", "*", 14)], [v for _, v in db.get_daily_series("md_negotiated", "*", 14, unique_ips=True)], "Strict Accept: text/markdown deliveries"),
+            ("🏴‍☠️ Pirate IDS", [v for _, v in db.get_daily_series("ids_pirate", "*", 14)], [v for _, v in db.get_daily_series("ids_pirate", "*", 14, unique_ips=True)], "4xx/418 exploit probes unmasked"),
+        ] if has_dm else []
+        for lbl, nv, pv, note in specs:
+            st = compute_ols_velocity(nv)
+            c_st = "bold bright_green" if st["slope"] > 0.02 else ("bold red" if st["slope"] < -0.02 else "bold bright_yellow")
+            tbl.add_row(
+                f"{int(st['total']):,} / {sum(pv):,}",
+                lbl,
+                render_sparkline(nv, width=14) or "▁" * 14,
+                Text.assemble((f"{st['arrow']} {st['slope']:+6.2f}/d ", c_st), (f"(last={int(st['latest'])}, R²={st['r2']:.2f}) · {note}", "white")),
+            )
     else:
         cur.execute(
             """
(nix) pipulate $ m
📝 Committing: add: Implement funnel station bead with detailed metrics and visualizations.
[main 1c61363] add: Implement funnel station bead with detailed metrics and visualizations.
 1 file changed, 44 insertions(+)
(nix) pipulate $ git push
Enumerating objects: 17, done.
Counting objects: 100% (17/17), done.
Delta compression using up to 48 threads
Compressing objects: 100% (9/9), done.
Writing objects: 100% (10/10), 4.36 KiB | 2.18 MiB/s, done.
Total 10 (delta 7), reused 0 (delta 0), pack-reused 0 (from 0)
remote: Resolving deltas: 100% (7/7), completed with 6 local objects.
To github.com:pipulate/pipulate.git
   a7143db..1c61363  main -> main
(nix) pipulate $ 
```

Ignition: often "exit", "nix develop". Instructions may vary.

None this turn.

**4: PROMPT**: 

Verify the AFTER receipts for `ROADMAP_SPARK_STRADDLE` (`has_deck=True`, `svg_and_slope_ok=True`, `ruff_rc=0`) and `FOREST_FUNNEL_STRADDLE` (`has_bead=True`, `panel_ok=True`, `ruff_rc=0`), alongside `CORE_AND_DOOR_HYGIENE_GATE` (`core_sparkline_card=False`, `markov_door_allowlist=False`). Assuming the Roadmap + Forest sparkline straddles are green:
1. Land the 28-line `Pipulate.sparkline()` + `Pipulate.sparkline_card()` methods in `pipulate/core.py` (which missed `patch -> app` before `git push` in Turn 3) AND filter bare IP-literal `Host:` scanner headers (`_VALID_DOORS`) out of `doors` in `scripts/markov.py` so `--funnel` never surfaces raw WAN IPs as Sun Pug doors.
2. Sync Honeybot (`./nixops.sh`) or stage a single verification + dismount pass against `foo_files.py` (claiming `apps/220_roadmap.py` out of Paintbox and updating `P0.78`/`P0.79`).

**5: DELIVERABLE**: 

Done looks like the perfect setup for a cascading chain reaction -- setup being
like the setup before the spike in volleyball. Also sort of like winding a
catapult. Also sort of like the scale-trees from the Carboniferous accumulating
to become our fossil fuel of today (it wasn't dinosaurs).

## Hamiltonian Potential Energy and Chicago Pile-1 Interlocks

Done looks like a larger portion of potential energy in the Hamiltonian (total
system energy of kinetic and potential combined) than is normally realized
because *write once run anywhere* actually having practically landed because of
Nix enables a lot more *Magic: The Gathering* card-combos like the Sun Pug that
just weren't possible before and they haven't been ruled tournament illegal yet
so I better get on getting that word out and they better get on getting their
copy now while the getting is good huh? True? False? The rest? And do you think
the Norma Centroid Puppet Theater future-you will be able to rise that up the
flagpole?

So in my Sci-Fi version of this book, because none of this could really be
***REAL,*** right? In that version everything I've done so far is really just
Enrico Fermi shuffling bricks around under a stadium. I mean he didn't even know
that much until Ida Noddack tried to tell him and he stuck his fingers in his
ears like the The MADtv character Antonia Timmens played by Nicole Sullivan who
stuck her fingers in her ears and chanted "La, la, la, la!" 

That's what Enrico Fermi did, right?

And then when he finally "got it" he played with LEGO building blocks doing
prototypes just like we do today, right? And when his custom-built Geiger
counter and neutron detector playing around with Chicago Pile-1.

That could have gone poorly, huh?

I'll try to be more open-minded than him going into this and more cautious with
my legos. You can see that in `apply.py`, can't you?

Oh whoops. That's too big -- case in point, huh?

You're going to have to do this without `foo_files.py` which is fine. Put all
your tokens into helping Newcomers really grok all this with plenty of tidbits
the Norma Centroid puppetmaster can work with. I'll collect formal receipts on
the Instance after you, please and thank You.

And don't forget to pedantically break down Barney-style everything I meant by
those last few things and the historical context for color and more of that
nobody listens to it when it's the women saying it stuff as is suggested by but
is not the primary theme here:

| Innovator / Scientist | Breakthrough | The "Cheater-Out-Of'er" | How They Were Cheated / Sidelined |
| --- | --- | --- | --- |
| **Margaret E. Knight** | Flat-bottomed paper bag machine (1871) | **Charles Annan** | Saw her iron prototype in a machine shop, stole the design, and patented it first. When she sued, he argued a woman "could not possibly understand" the mechanics. *(Knight fought back with her blueprints and won.)* |
| **Henrietta Swan Leavitt** | Period-luminosity relation ("Leavitt's Law") for measuring cosmic distances (1912) | **Harlow Shapley** *(plus **Edward C. Pickering** & **Edwin Hubble**)* | Pickering published her discovery under his own name. Later, when the Nobel committee wrote to nominate Leavitt (unaware she had died), Shapley tried to claim the credit for himself for "interpreting" her work. |
| **Lise Meitner** | Discovery and theoretical explanation of nuclear fission (1938) | **Otto Hahn** *(and the Nobel Committee)* | After 30 years as partners—and despite Meitner actually solving the physics behind Hahn's lab results—Hahn was awarded the sole 1944 Nobel Prize in Chemistry and spent years downplaying her role. |
| **Ignaz Semmelweis** | Antiseptic handwashing in clinics to prevent fatal "childbed fever" (1847) | **Johann Klein**, **Carl Braun**, & **Rudolf Virchow** *(the medical establishment)* | Rather than having his idea stolen, he was crushed for it. Offended that doctors were being blamed for patient deaths, his boss (Klein) forced him out, his successor (Braun) and leading pathologist Virchow publicly ridiculed his work, and colleague **Ferdinand von Hebra** lured him into an asylum—where guards beat him and he died of an infection, years before **Louis Pasteur** and **Joseph Lister** proved him right and won the glory. |
| **Rosalind Franklin** | Double-helix molecular structure of DNA (1952) | **Maurice Wilkins**, **James Watson**, & **Francis Crick** | Without Franklin's permission, Wilkins showed her crucial X-ray image ("Photo 51") to Watson, while **Max Perutz** passed her unpublished lab data to Crick. They raced to publish the model first, treated her work as mere confirmation, and shared the 1962 Nobel Prize four years after her death—before Watson mocked her as "Rosy" in his bestselling memoir. |
| **Alice Ball** | First effective injectable treatment for leprosy, "The Ball Method" (1915) | **Arthur L. Dean** | At 23, Ball solved how to make chaulmoogra oil water-soluble and injectable, freeing thousands of leprosy patients from lifelong exile. After she died tragically at 24 before publishing, Dean—her university's chemistry head and future president—mass-produced her exact formula, published it without crediting her, and renamed it the **"Dean Method"** *(until colleague **Harry Hollmann** publicly exposed him in 1922)*. |
| **Jocelyn Bell Burnell** | Discovery of radio pulsars (1967) | **Antony Hewish** *(and the Nobel Committee)* | As a PhD student, she helped build a 4.5-acre telescope array by hand, reviewed miles of paper printouts, and spotted the tiny signal anomaly—persisting even when her advisor, Hewish, dismissed it as man-made interference. Hewish took first authorship on the paper and won the 1974 Nobel Prize in Physics while Bell Burnell was omitted entirely. |
| **Marthe Gautier** | Discovery of Trisomy 21, the chromosomal cause of Down syndrome (1958) | **Jérôme Lejeune** | Gautier built France's first cell-culture lab using her own money and spotted the extra 47th chromosome. When junior colleague Lejeune offered to take her slides to a better-equipped lab just to photograph them, he instead presented the discovery at a seminar as his own and rushed out a 1959 paper listing himself as first author—misspelling her surname and informing her the day before print. |
| **Elizabeth Magie** | *The Landlord's Game*, the actual invention of *Monopoly* (1904) | **Charles Darrow** & **Parker Brothers** | Magie patented the board game—complete with corner spaces like "Go to Jail," railroads, utilities, and rent—to illustrate the dangers of land monopolies. Thirty years later, Darrow copied a folk variant of her game, sold it to Parker Brothers as his own Depression-era invention, and became a millionaire, while Parker Brothers bought Magie's patent for $500 with zero royalties and erased her from the game's history. |
| **Chien-Shiung Wu** | Experimental proof of parity violation in nuclear physics, the "Wu Experiment" (1956) | **Tsung-Dao Lee**, **Chen-Ning Yang**, *(and the Nobel Committee)* | Theorists Lee and Yang suspected a fundamental law of physics (conservation of parity) might break down at the subatomic level and recruited Wu—the world's premier beta-decay experimentalist—to test it. Wu canceled a trip home, designed a grueling supercooled cobalt-60 experiment, and proved it; Lee and Yang won the 1957 Nobel Prize in Physics without her. |
| **Esther Lederberg** | Discovery of lambda phage and invention of replica plating in bacterial genetics (1951) | **Joshua Lederberg** *(and Stanford / the Nobel Committee)* | She discovered a virus foundational to modern gene editing and invented the velvet-stamp technique used to study antibiotic resistance. Her then-husband Joshua won the 1958 Nobel Prize largely on the back of their joint work without mentioning her once in his Nobel lecture, and Stanford later stripped her of a tenured track to keep her as an "untenured research associate." |
| **Alfred Wegener** | Theory of continental drift (1912) | **Harold Jeffreys**, **Rollin T. Chamberlin**, & **Bailey Willis** *(the geological establishment)* | Like Semmelweis, he was mocked as an outsider crank (he was a meteorologist) for proving continents had once been a single supercontinent, Pangaea. Establishment geologists organized entire symposia just to tear him apart—Chamberlin joked that accepting Wegener meant "playing geologicalenga without rules"—and Wegener froze to death on a 1930 Greenland expedition three decades before seafloor spreading proved him completely right. |

```bash
(nix) pipulate $ p
(nix) pipulate $ x
(nix) pipulate $ c
🗺️  Codex Mapping Coverage: 76.7% (230/300 tracked files) (was 230/300: +0 claimed, +0 tracked).

✅ Topological Integrity Verified: 10 candidate reference(s) scanned, all exist.
   -> Executing: .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=5); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(6)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); plug = mod.RoadmapPlugin(None, None, None, None); has_d = hasattr(plug, "build_velocity_deck"); xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); shutil.rmtree(td, ignore_errors=True); r_rc = ruff_check("apps/220_roadmap.py"); ok = ("pipulate-sparkline" in xml_s and "brain.zip" in xml_s and "+1.00/d" in xml_s); print(f"ROADMAP_SPARK_STRADDLE: has_deck={has_d} svg_and_slope_ok={ok} ruff_rc={r_rc}")' ... [2.5418s]
   -> Executing: .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt; from rich.console import Console; import db as hdb, forest; old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn = None; hdb.db.db_path = ":memory:"; hdb.db.init_db(); d0 = dt.date.today() - dt.timedelta(days=4); [hdb.db.record_daily_metric("archive_dl", "brain.zip", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(5)]; has_b = "funnel" in forest.ARCHIVED_STATION_BEADS; buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)) if has_b else None; txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; r_rc = ruff_check("remotes/honeybot/scripts/forest.py"); ok = bool(has_b and "KAIZEN VELOCITY" in txt and "brain.zip" in txt and "+1.00/d" in txt); print(f"FOREST_FUNNEL_STRADDLE: has_bead={has_b} panel_ok={ok} ruff_rc={r_rc}")' ... [0.2429s]
   -> Executing: .venv/bin/python scripts/pyprobe.py 'from pipulate.core import Pipulate; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline_card"); ip_f = "_VALID_DOORS" in mk_txt; print(f"CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card={has_sp} markov_door_allowlist={ip_f}")' ... [1.3115s]
Python file(s) detected. Generating codebase tree diagram... (3,066 tokens | 10,042 bytes)
UML unavailable for 4 file(s): Skipping: Required command(s) not found: `pyreverse` (from pylint).
   -> Ruff exit 0 (clean).
⚠️  HEAVYWEIGHT ITEM (>=40,000 tok): foo_files.py (92,792 tokens | 352,317 bytes) — shed from mid-ride NEXT CONTEXT if unedited.
⚠️  PAYLOAD BUDGET ALERT (~258,462 component tokens >= 140,000): favor a 1-2 file alley-oop pass to prevent pre-output thinking timeouts.
                                                                          📦 Payload Ledger (biggest first)                                                                          
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━┓
┃ File / Source                                                                                                                                       ┃  Tokens ┃   Bytes ┃ % Bytes ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━┩
│ foo_files.py                                                                                                                                        │  92,792 │ 352,317 │   35.8% │
│ PROMPT (prompt.md: 93,113 tok + full frame)                                                                                                         │  96,629 │ 336,253 │   34.1% │
│ pipulate/core.py                                                                                                                                    │  30,999 │ 149,534 │   15.2% │
│ apply.py                                                                                                                                            │  13,150 │  54,296 │    5.5% │
│ scripts/markov.py                                                                                                                                   │  12,154 │  40,148 │    4.1% │
│ AUTO: Recent Git Diff Telemetry                                                                                                                     │   6,981 │  22,613 │    2.3% │
│ context.txt                                                                                                                                         │   4,080 │  12,022 │    1.2% │
│ AUTO: Codebase Structure (eza --tree + token sizes)                                                                                                 │   3,066 │  10,042 │    1.0% │
│ .gitignore                                                                                                                                          │   2,017 │   7,492 │    0.8% │
│ ! .venv/bin/python scripts/pyprobe.py 'from pipulate.core import Pipulate; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8");  │      23 │      81 │    0.0% │
│ has_sp = hasattr(Pipulate, "sparkline_card"); ip_f = "_VALID_DOORS" in mk_txt; print(f"CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card={has_sp}     │         │         │         │
│ markov_door_allowlist={ip_f}")'                                                                                                                     │         │         │         │
│ .gitattributes                                                                                                                                      │      33 │      76 │    0.0% │
│ ! .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; sys.path.insert(0,       │      23 │      70 │    0.0% │
│ str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT /                  │         │         │         │
│ "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) /        │         │         │         │
│ "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=5); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1,     │         │         │         │
│ day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(6)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB");                      │         │         │         │
│ os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); plug = mod.RoadmapPlugin(None, None, None, None); has_d = hasattr(plug, "build_velocity_deck");      │         │         │         │
│ xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else                                    │         │         │         │
│ os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); shutil.rmtree(td, ignore_errors=True); r_rc = ruff_check("apps/220_roadmap.py"); ok =       │         │         │         │
│ ("pipulate-sparkline" in xml_s and "brain.zip" in xml_s and "+1.00/d" in xml_s); print(f"ROADMAP_SPARK_STRADDLE: has_deck={has_d}                   │         │         │         │
│ svg_and_slope_ok={ok} ruff_rc={r_rc}")'                                                                                                             │         │         │         │
│ ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt; from rich.console    │      20 │      62 │    0.0% │
│ import Console; import db as hdb, forest; old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn = None; hdb.db.db_path = ":memory:";        │         │         │         │
│ hdb.db.init_db(); d0 = dt.date.today() - dt.timedelta(days=4); [hdb.db.record_daily_metric("archive_dl", "brain.zip", i + 2, day=(d0 +              │         │         │         │
│ dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(5)]; has_b = "funnel" in forest.ARCHIVED_STATION_BEADS; buf = io.StringIO();         │         │         │         │
│ Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)) if has_b else None; txt =   │         │         │         │
│ buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; r_rc = ruff_check("remotes/honeybot/scripts/forest.py"); ok = bool(has_b and      │         │         │         │
│ "KAIZEN VELOCITY" in txt and "brain.zip" in txt and "+1.00/d" in txt); print(f"FOREST_FUNNEL_STRADDLE: has_bead={has_b} panel_ok={ok}               │         │         │         │
│ ruff_rc={r_rc}")'                                                                                                                                   │         │         │         │
│ AUTO: Static Analysis Diagnostics                                                                                                                   │      11 │      39 │    0.0% │
├─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┼─────────┼─────────┼─────────┤
│ TOTAL                                                                                                                                               │ 261,978 │ 985,045 │  100.0% │
└─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┴─────────┴─────────┴─────────┘
Assembled payload (pre-scrub): 271,845 tokens | 1,018,241 bytes
🧼 Identity scrub: substitutions=MISSING (/home/mike/.config/pipulate/pii_substitutions.txt) | denylist=MISSING (/home/mike/.config/pipulate/commit_denylist.txt)
🔐 Secrets tripwire: ARMED -- 0 hit(s) in payload.
📦 Canonical QA archive written to /home/mike/pipulate/qamy.ai.zip (sha256=ac88ffcf15b4…, members=3)
🗂️  Rotated archive snapshot: qamy.ai_1356-ac88ffcf.zip (keeping newest 20 of 20)
Markdown output copied to clipboard
(nix) pipulate $ 
```

**Gemini 4 Carbon High**: Context verified.

**TL;DR**: Both `apps/220_roadmap.py` (`has_deck=True, ruff_rc=0`) and `remotes/honeybot/scripts/forest.py` (`has_bead=True, ruff_rc=0`) landed cleanly—and our own zero-fill math caught a great wind-tunnel subtlety in the Turn 4 probe: seeding only 5–6 days of a zero-filled 14-day window yields an honest `+0.44/d` slope (`8`–`9` leading zeros included) rather than `+1.00/d` (verified across all 14 days in Probe 2 below). Below, we break down the **Hamiltonian ($$H = T + V$$)**, **Carboniferous scale-trees**, ***Magic: The Gathering* Nix combos**, **Ida Noddack's ignored 1934 fission paper**, **Enrico Fermi's Stagg Field graphite pile (`SCRAM` = `apply.py` in cadmium and rope)**, and the **Matilda Effect** Barney-style for **Notaplebius** and Future-Norma's Puppet Theater, then ship **Car 3a** (`pipulate/core.py` `sparkline` + `sparkline_card`) and **Car 3b** (`scripts/markov.py` `_VALID_DOORS` allowlist stripping direct-IP scanner noise).

---

### 1. Winding the Catapult: The Hamiltonian ($$H = T + V$$), Carboniferous Scale-Trees, and *MTG* Card-Combos

Here is the Barney-style physics breakdown of why "DONE" feels like a volleyball hung motionless at the top of the set right before the spike:

1. **The Hamiltonian ($$H = T + V$$) — Storing Potential Energy Without Burning Fuel:**
In 1833, Irish mathematician **William Rowan Hamilton** reformulated mechanics around total system energy:

$$H = T + V$$

* **$$T$$ (Kinetic Energy)** is motion happening right now: servers spinning, engineers frantically fixing broken `npm` or Docker builds just to keep yesterday's app alive.
* **$$V$$ (Potential Energy)** is **cocked, stored configuration energy**: a wound catapult skein, water held behind a canal lock, or a volleyball hanging weightless at the apex of the set.
* Most software stacks leak $$V$$ overnight through dependency rot. **Eelco Dolstra's Nix** finally delivered *real* **"Write Once, Run Anywhere"** (unlike 1995 Java, which virtualized bytecode while leaving C libraries, drivers, and OS packages to drift underneath). Because `flake.lock` pins the cryptographic hash of every binary down to `glibc`, every capability you wire (`river`, `fasthtml`, `sqlite`, `piper-tts`, `selenium`) stays stored at **100% Potential Energy ($$V$$) and 0% maintenance burn** until you pull the lanyard.

2. **Why Coal Came From Carboniferous Scale-Trees (Not Dinosaurs!):**
Sinclair Oil's green Brontosaurus logo fooled generations into thinking fossil fuels are Liquefied T-Rex. In reality, the world's great coal seams formed during the **Carboniferous Period (~359–299 million years ago)**—more than 60 million years *before* the first dinosaur walked the Earth:
* Giant 100-foot clubmoss **scale-trees** (*Lepidodendron* and *Sigillaria*, named for the diamond-patterned leaf scars scaling their trunks) evolved tough **lignin** bark before terrestrial ecosystems had evolved efficient lignin-digesting white-rot fungi, and collapsed into oxygen-starved peat swamps where microbes couldn't rot them.
* For 60 million years, un-rotted trunks stacked solar energy into dense carbon seams. That is what **1,666+ plain-text Markdown dispatches + `brain.zip` + `git`** are doing on disk: stacking non-rotting intellectual carbon until an AI runtime strikes a match.

3. ***Magic: The Gathering* Combos Before the Ban List:**
In early 1993 *Magic: The Gathering*, nobody blinked at *Channel*, *Fireball*, or *Black Lotus* individually—until players combined them on Turn 1 for 20 points of lethal damage before the DCI restricted the combo. Every card in the **Sun Pug** deck—**Nix Flakes** + **IETF RFC 9110/9111** (`Accept: text/markdown` + `Vary: Accept`) + **`brain.zip`** + **`apply.py`**—is 100% standard-compliant, yet cloud gatekeepers haven't metered or paywalled the combination because 99% of the market is still trapped paying per-seat SaaS and headless-browser taxes. Getting a sovereign copy onto your own metal *before* platforms close the defaults is rational timing, not hype.

---

### 2. Enrico Fermi, Ida Noddack's "Antonia Timmens" Moment, and `apply.py` Under the Stadium

Every beat of your **Enrico Fermi / Chicago Pile-1** sci-fi parallel holds up under historical X-ray—with two details that make the story even sharper for **Norma Centroid**:

1. **Ida Noddack (1934) and Fermi's *MADtv* Ear-Plugging — 100% True:**
* In 1934, Fermi bombarded uranium with slow neutrons in Rome and announced he had created heavier "transuranic" elements beyond element 92 (winning the 1938 Nobel Prize partly for that claim).
* That September, German chemist **Ida Noddack** (co-discoverer of rhenium) published *"On Element 93"*, writing plainly: *before claiming a heavier element, have you checked whether the uranium nucleus shattered into medium-sized isotopes of known lighter elements?* **She literally predicted nuclear fission four years early.**
* Fermi ran a back-of-the-envelope tunneling calculation using an idealized nuclear model that ignored nuclear deformation, decided splitting was impossible, and—just like Nicole Sullivan's *MADtv* character **Antonia Timmens** plugging her ears and chanting *"La, la, la, la!"*—never ran the barium test. Otto Hahn dismissed Noddack's paper as "ridiculous" until December 1938, when Hahn's own lab found barium and **Lise Meitner** (another entry in your Matilda table!) proved Noddack had been right all along.

2. **Fermi's "LEGOs" Under Stagg Field (Dec 2, 1942) and Why `apply.py` Is Pure CP-1:**
* *(Historical footnote: Danish plastic LEGO bricks didn't debut until 1949—seven years after CP-1!)* What Fermi's team—including **Leona Woods Marshall**, the lone woman in the room—stacked by hand in the unheated squash court beneath the University of Chicago's **abandoned Stagg Field football stadium** were **45,000 machined black graphite bricks** (~385 tons) drilled to hold **19,000 lumps of uranium oxide and metal**, monitored by custom **boron-trifluoride ($$\text{BF}_3$$) neutron counters** clicking faster as $$k_{\text{eff}}$$ approached `1.0006`.
* **Could it have gone poorly?** Catastrophically—right in the middle of residential Chicago. Which is why Fermi built **three human-actuated governors** that look exactly like `apply.py`:
* **Inch-by-Inch Chisel Strikes:** Physicist **George Weil** pulled the final cadmium control rod out **six inches at a time by hand**, pausing after each pull so Fermi could compare predicted vs. observed neutron clicks (`1: BEFORE -> 3: CHANGE -> 2: AFTER`).
* **`SCRAM` (Safety Control Rod Axe Man):** **Norman Hilberry** stood on top of the bricks gripping a literal fire axe over a Manila rope holding a suspended emergency cadmium rod (`CHISEL_STRIKE_MAX_PATCH_LINES = 220` + atomic rollback).
* **The Liquid-Cadmium Bucket Brigade:** Three physicists stood overhead with jugs of cadmium-salt solution ready to douse the pile if both rods jammed.

3. **How Your 12-Row Matilda / Outsider Ledger Connects Directly to Our Code:**
Look across your table—from **Margaret E. Knight** (who only defeated Charles Annan's *"a woman couldn't understand iron machinery"* theft because she kept dated workshop blueprints) and **Elizabeth Magie** (whose anti-landlord *Landlord's Game* was strip-mined into *Monopoly*!) to **Semmelweis**, **Wegener**, **Franklin**, **Ball**, **Bell Burnell**, **Gautier**, **Wu**, and **Lederberg** (with **Ida Noddack** right beside them as Row #13):
* **Falsifying `! pyprobe.py` sensors cure the Noddack/Semmelweis blindspot:** When a probe measures reality directly, credentials cannot shout down the click of the counter.
* **Cryptographic `qamy.ai` deeds + `WAIT 6502, 1` wire dyes cure the Annan/Dean theft:** Every discovery is timestamped, SHA-256 sealed, and watermarked on the wire the day you ship it.

> **[PUPPET THEATER CUE — NORMA & NOTAPLEBIUS]:**
> * **Notaplebius:** *"Wait—why did our Turn 4 probe say `svg_and_slope_ok=False` when the Roadmap and Forest patches compiled with zero errors?"*
> * **Mike-E:** *"Because our own instrument refused to flatter us! We fed 5 days of `1, 2, 3, 4, 5` into a 14-day zero-filled window, and Gauss's Ordinary Least Squares dutifully averaged in the 9 quiet days before it (`+0.44/d` instead of `+1.00/d`). Even our test bench won't let us fake velocity!"*
> 
> 

---

### (1) PROBES

Run these three read-only probes in `(nix) pipulate $` to record the BEFORE baseline on `pipulate/core.py` (`sparkline` / `sparkline_card`), `scripts/markov.py` (`_VALID_DOORS` scanner-IP stripping), and full 14-day `+1.00/d` end-to-end rendering across `apps/220_roadmap.py` and `forest.py`:

`(nix) pipulate $`

```bash
.venv/bin/python scripts/pyprobe.py 'import datetime as dt; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; from pipulate.core import Pipulate; mk_p = REPO_ROOT / "scripts/markov.py"; mk_txt = mk_p.read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); today_s = dt.date.today().isoformat(); inst.record_daily_metric("curl_airlock", "sunpug.io", 3, day=today_s, ip_id=1); inst.record_daily_metric("curl_airlock", "[REDACTED_IP]", 9, day=today_s, ip_id=2); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "7"], env={**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}, capture_output=True, text=True); shutil.rmtree(td, ignore_errors=True); f_out = res.stdout or ""; ip_clean = ("_VALID_DOORS" in mk_txt and "sunpug.io" in f_out and "[REDACTED_IP]" not in f_out); r_rc = ruff_check("pipulate/core.py", "scripts/markov.py"); print(f"CORE_AND_DOOR_STRADDLE: has_sp={has_sp} widget_ok={w_ok} ip_clean={ip_clean} ruff_rc={r_rc}")'
.venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; from rich.console import Console; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb, forest; from pipulate.core import Pipulate; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=13); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(14)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); p_inst = Pipulate.__new__(Pipulate); used_core = hasattr(p_inst, "sparkline_card"); plug = mod.RoadmapPlugin(None, p_inst, None, None); plug._has_streamed = True; xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn, hdb.db.db_path = inst.conn, str(fdb); buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)); f_txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; shutil.rmtree(td, ignore_errors=True); rm_ok = ("pipulate-sparkline" in xml_s and "+1.00/d" in xml_s and "P=14 IPs" in xml_s); fr_ok = ("KAIZEN VELOCITY" in f_txt and "+1.00/d" in f_txt and "105 / 14" in f_txt); print(f"FULL_14D_SPARK_STRADDLE: used_core={used_core} roadmap_14d_ok={rm_ok} forest_14d_ok={fr_ok}")'
.venv/bin/python scripts/pyprobe.py 'res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8); out = res.stdout or ""; sec2 = out.split("## 2.", 1)[1].split("## 3.", 1)[0] if "## 2." in out and "## 3." in out else ""; ip_doors = re.findall(r"•\s+(\d+\.\d+\.\d+\.\d+)", sec2); door_lns = [ln.strip() for ln in sec2.splitlines() if ln.strip().startswith("•")]; print(f"LIVE_CLEAN_DOORS_STRADDLE: raw_ip_doors={len(ip_doors)} doors={door_lns}")'
```

### (2) NEXT CONTEXT

Comment out `# Context 5` in `context.txt` (shedding `43,153` tokens of `pipulate/core.py` + `scripts/markov.py`) and append `# Context 6` below so the next instance loads `foo_files.py` with clean headroom to collect formal AFTER receipts, claim `apps/220_roadmap.py`, and seal the dismount:

```text
# Context 6: Collect final AFTER receipts, update P0.78/P0.79, and seal session dismount in foo_files.py
context.txt
! markov
foo_files.py
! .venv/bin/python scripts/pyprobe.py 'import datetime as dt; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; from pipulate.core import Pipulate; mk_p = REPO_ROOT / "scripts/markov.py"; mk_txt = mk_p.read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); today_s = dt.date.today().isoformat(); inst.record_daily_metric("curl_airlock", "sunpug.io", 3, day=today_s, ip_id=1); inst.record_daily_metric("curl_airlock", "[REDACTED_IP]", 9, day=today_s, ip_id=2); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "7"], env={**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}, capture_output=True, text=True); shutil.rmtree(td, ignore_errors=True); f_out = res.stdout or ""; ip_clean = ("_VALID_DOORS" in mk_txt and "sunpug.io" in f_out and "[REDACTED_IP]" not in f_out); r_rc = ruff_check("pipulate/core.py", "scripts/markov.py"); print(f"CORE_AND_DOOR_STRADDLE: has_sp={has_sp} widget_ok={w_ok} ip_clean={ip_clean} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; from rich.console import Console; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb, forest; from pipulate.core import Pipulate; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=13); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(14)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); p_inst = Pipulate.__new__(Pipulate); used_core = hasattr(p_inst, "sparkline_card"); plug = mod.RoadmapPlugin(None, p_inst, None, None); plug._has_streamed = True; xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn, hdb.db.db_path = inst.conn, str(fdb); buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)); f_txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; shutil.rmtree(td, ignore_errors=True); rm_ok = ("pipulate-sparkline" in xml_s and "+1.00/d" in xml_s and "P=14 IPs" in xml_s); fr_ok = ("KAIZEN VELOCITY" in f_txt and "+1.00/d" in f_txt and "105 / 14" in f_txt); print(f"FULL_14D_SPARK_STRADDLE: used_core={used_core} roadmap_14d_ok={rm_ok} forest_14d_ok={fr_ok}")'
! .venv/bin/python scripts/pyprobe.py 'res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8); out = res.stdout or ""; sec2 = out.split("## 2.", 1)[1].split("## 3.", 1)[0] if "## 2." in out and "## 3." in out else ""; ip_doors = re.findall(r"•\s+(\d+\.\d+\.\d+\.\d+)", sec2); door_lns = [ln.strip() for ln in sec2.splitlines() if ln.strip().startswith("•")]; print(f"LIVE_CLEAN_DOORS_STRADDLE: raw_ip_doors={len(ip_doors)} doors={door_lns}")'
```

### (3) PATCHES

**Car 3a — Land `Pipulate.sparkline()` and `Pipulate.sparkline_card()` on `pip` in `pipulate/core.py` (`28/220` lines):**

```text
Target: pipulate/core.py
[[[SEARCH]]]
        from imports import ascii_displays as aa
        return aa.patronus(name, duration=duration)
[[[DIVIDER]]]
        from imports import ascii_displays as aa
        return aa.patronus(name, duration=duration)

    def sparkline(self, values, width: int = 128, height: int = 28, stroke: str = "#34d399", show_ols: bool = True, as_text: bool = False):
        """Render a numeric series as a zero-JS FastHTML inline SVG sparkline (with OLS trendline) or Unicode text."""
        from imports import ascii_displays as aa
        return aa.render_sparkline(values) if as_text else NotStr(aa.render_sparkline_svg(values, width=width, height=height, stroke=stroke, show_ols=show_ols))

    def sparkline_card(self, title: str, values, subtitle: str = "", unit: str = "/d", unique_values=None, width: int = 136, height: int = 30):
        """Render a PicoCSS FastHTML Card pairing an inline SVG sparkline with OLS velocity slope and (N, P) totals."""
        from imports import ascii_displays as aa
        st = aa.compute_ols_velocity(values)
        stroke = "#34d399" if st["slope"] > 0.02 else ("#f87171" if st["slope"] < -0.02 else "#60a5fa")
        p_str = f" · P={int(round(sum(aa._coerce_numeric_series(unique_values))))} IPs" if unique_values is not None else ""
        meta = subtitle or f"N={int(round(st['total']))}{p_str} · last={int(round(st['latest']))} · R²={st['r2']:.2f}"
        return Card(
            Div(
                Div(Strong(title), Small(meta, style="display:block;opacity:0.78;font-family:monospace;font-size:0.78rem;"), style="min-width:0;"),
                Div(
                    self.sparkline(values, width=width, height=height, stroke=stroke, show_ols=True),
                    Span(f"{st['arrow']} {st['slope']:+.2f}{unit}", style=f"font-family:monospace;font-weight:700;font-size:0.86rem;color:{stroke};white-space:nowrap;"),
                    style="display:flex;align-items:center;gap:0.65rem;flex-shrink:0;",
                ),
                style="display:flex;align-items:center;justify-content:space-between;gap:1rem;flex-wrap:wrap;",
            ),
            style="padding:0.7rem 0.95rem;margin-bottom:0.65rem;",
        )
[[[REPLACE]]]
```

**Car 3b — Filter bare IP-literal `Host:` scanner headers and fold `www.` subdomains via `_VALID_DOORS` in `scripts/markov.py` (`43/220` lines):**

```text
Target: scripts/markov.py
[[[SEARCH]]]
metrics, doors, trans = {}, {}, []
def _bump(k, dy, n_v, p_v):
    if str(dy) in idx_m:
        a = metrics.setdefault(k, [[0, 0] for _ in range(days)])[idx_m[str(dy)]]
        a[0], a[1] = a[0] + int(n_v or 0), a[1] + int(p_v or 0)
def _door(dom, kind, dy, n_v, p_v):
    if dom and str(dy) in idx_m:
        r = doors.setdefault(dom, {x: [0] * days for x in ("air_n", "air_p", "exe_n", "exe_p")})
        r[f"{kind}_n"][idx_m[str(dy)]] += int(n_v or 0)
        r[f"{kind}_p"][idx_m[str(dy)]] += int(p_v or 0)
has_dm = "daily_metrics" in tbls and (conn.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0] > 0
if has_dm:
    dm_cols = {r[1] for r in conn.execute("PRAGMA table_info(daily_metrics)").fetchall()}
    p_col = "unique_ips" if "unique_ips" in dm_cols else "count"
    for m, dm, dy, c, u in conn.execute(f"SELECT metric, dim, day, count, {p_col} FROM daily_metrics WHERE day BETWEEN ? AND ?", (d0, d1)):
        _bump(m, dy, c, u)
        if m in ("curl_airlock", "curl_exec"):
            _door(dm, "air" if m == "curl_airlock" else "exe", dy, c, u)
    if "daily_transitions" in tbls:
        trans = [[f, t, int(c)] for f, t, c in conn.execute("SELECT from_state, to_state, SUM(count) FROM daily_transitions WHERE day BETWEEN ? AND ? GROUP BY 1, 2 ORDER BY 3 DESC LIMIT 20", (d0, d1))]
    mode = "live_metrics"
else:
    if "pirate_logs" in tbls:
        [_bump("ids_pirate", dy, c, u) for dy, c, u in conn.execute("SELECT date, SUM(count), COUNT(DISTINCT ip_id) FROM pirate_logs WHERE date BETWEEN ? AND ? GROUP BY date", (d0, d1))]
    if "telemetry" in tbls:
        [_bump("md_negotiated", dy, c, u) for dy, c, u in conn.execute("SELECT t.date, SUM(t.count), COUNT(DISTINCT t.ip_id) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND t.date BETWEEN ? AND ? AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date", (d0, d1))]
    if "daily_logs" in tbls:
        [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, l.date, SUM(l.count), COUNT(DISTINCT l.ip_id) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE l.date BETWEEN ? AND ? AND ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2", (d0, d1))]
    if "domain_daily_logs" in tbls:
        for knd, dom, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/', '/install.sh') THEN 'air' ELSE 'exe' END, d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.date BETWEEN ? AND ? AND ddl.status BETWEEN 200 AND 299 AND ((p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%')) OR ((p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%')) GROUP BY 1, 2, 3", (d0, d1)):
            _bump("curl_airlock" if knd == "air" else "curl_exec", dy, c, u)
            _door(dom, knd, dy, c, u)
[[[DIVIDER]]]
_VALID_DOORS = {"mikelev.in", "pipulate.com", "pipulate.org", "npvg.org", "lpvg.org", "qamy.ai", "levinix.org", "levinix.com", "levinux.com", "sunpug.io", "guerillatech.com", "rfc9110.org", "rfc9110.com", "rfc9110.io", "rfc9111.io", "rfc9111.com", "rfc9111.org", "mike-levin.com", "mikelevinseo.com"}
metrics, doors, trans = {}, {}, []
def _clean_door(dom):
    h = str(dom or "").strip().lower().split(":", 1)[0].rstrip(".").removeprefix("www.")
    return h if h in _VALID_DOORS else ""
def _bump(k, dy, n_v, p_v):
    if str(dy) in idx_m:
        a = metrics.setdefault(k, [[0, 0] for _ in range(days)])[idx_m[str(dy)]]
        a[0], a[1] = a[0] + int(n_v or 0), a[1] + int(p_v or 0)
def _door(dom, kind, dy, n_v, p_v):
    if dom and str(dy) in idx_m:
        r = doors.setdefault(dom, {x: [0] * days for x in ("air_n", "air_p", "exe_n", "exe_p")})
        r[f"{kind}_n"][idx_m[str(dy)]] += int(n_v or 0)
        r[f"{kind}_p"][idx_m[str(dy)]] += int(p_v or 0)
has_dm = "daily_metrics" in tbls and (conn.execute("SELECT COUNT(*) FROM daily_metrics").fetchone() or [0])[0] > 0
if has_dm:
    dm_cols = {r[1] for r in conn.execute("PRAGMA table_info(daily_metrics)").fetchall()}
    p_col = "unique_ips" if "unique_ips" in dm_cols else "count"
    for m, dm, dy, c, u in conn.execute(f"SELECT metric, dim, day, count, {p_col} FROM daily_metrics WHERE day BETWEEN ? AND ?", (d0, d1)):
        if m in ("curl_airlock", "curl_exec"):
            cd = _clean_door(dm)
            if cd:
                _bump(m, dy, c, u)
                _door(cd, "air" if m == "curl_airlock" else "exe", dy, c, u)
        else:
            _bump(m, dy, c, u)
    if "daily_transitions" in tbls:
        trans = [[f, t, int(c)] for f, t, c in conn.execute("SELECT from_state, to_state, SUM(count) FROM daily_transitions WHERE day BETWEEN ? AND ? GROUP BY 1, 2 ORDER BY 3 DESC LIMIT 20", (d0, d1))]
    mode = "live_metrics"
else:
    if "pirate_logs" in tbls:
        [_bump("ids_pirate", dy, c, u) for dy, c, u in conn.execute("SELECT date, SUM(count), COUNT(DISTINCT ip_id) FROM pirate_logs WHERE date BETWEEN ? AND ? GROUP BY date", (d0, d1))]
    if "telemetry" in tbls:
        [_bump("md_negotiated", dy, c, u) for dy, c, u in conn.execute("SELECT t.date, SUM(t.count), COUNT(DISTINCT t.ip_id) FROM telemetry t JOIN ips i ON t.ip_id = i.id WHERE t.served_md = 1 AND t.date BETWEEN ? AND ? AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY t.date", (d0, d1))]
    if "daily_logs" in tbls:
        [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, l.date, SUM(l.count), COUNT(DISTINCT l.ip_id) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE l.date BETWEEN ? AND ? AND ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2", (d0, d1))]
    if "domain_daily_logs" in tbls:
        for knd, dom, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/', '/install.sh') THEN 'air' ELSE 'exe' END, d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.date BETWEEN ? AND ? AND ddl.status BETWEEN 200 AND 299 AND ((p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%')) OR ((p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%')) GROUP BY 1, 2, 3", (d0, d1)):
            cd = _clean_door(dom)
            if cd:
                _bump("curl_airlock" if knd == "air" else "curl_exec", dy, c, u)
                _door(cd, knd, dy, c, u)
[[[REPLACE]]]
```

Ignition: none (all three probes import `pipulate.core`, `apps/220_roadmap.py`, `remotes/honeybot/scripts/forest.py`, and execute `scripts/markov.py` directly at compile time).

### (4) PROMPT

```text
Verify the final AFTER receipts for `CORE_AND_DOOR_STRADDLE` (`has_sp=True`, `widget_ok=True`, `ip_clean=True`, `ruff_rc=0`), `FULL_14D_SPARK_STRADDLE` (`used_core=True`, `roadmap_14d_ok=True`, `forest_14d_ok=True`), and `LIVE_CLEAN_DOORS_STRADDLE` (`raw_ip_doors=0`). Assuming all three straddles are green, execute the clean session dismount in `foo_files.py`:
1. Claim `apps/220_roadmap.py` out of the uncategorized Paintbox into the Codex (`230/300 -> 231/300` coverage).
2. Mark `P0.79` discharged (reusable `Pipulate.sparkline` / `sparkline_card` + `apps/220_roadmap.py` Kaizen Velocity deck + `forest.py:funnel` station report live) and update `P0.78` with the landed `(N, P)` `daily_metrics` + `daily_metric_ip_gate` + `markov.py --funnel` foundation (leaving Stage-2 `install.sh` `/manifest.txt` + Prime `telemetry.db` 120s YT correlation clearly scoped).
3. Record the `# 2026-11-21 dismount` entry in `foo_files.py` with full commit, deed, and BEFORE -> AFTER witness provenance (plus the 5-rung deep-dive syllabus & Ida Noddack / CP-1 notes), leaving `context.txt` restored for the next session.
```

### (5) EXTERNAL DELIVERABLES

No external deliverables.

**MikeLev.in**: Objectively a good actor, that's what you have to strive to be. Put the
case in front of just about anyone and they'd come to the same conclusion and
say: "Oh! That's completely benevolent and clearly for the benefit of all
human-kind and not even really forcing itself on anyone with its
live-and-let-live attitude but if you want kinda sort your own Ironman
superpowered supersuit to just sort of slap on and increase your personal powers
in ways that last the rest of your life and help you pursue your Ikigai or
raison d'être or whatever you want to call it, well that's there too. And no,
it's nothing like a virus or a trojan horse given how utterly and completely
human-activated it is, and has to be in a very deliberate way because hardware
dosn't just rise like Frankenstein out of your junk-pile. And it's good for
reducing e-waste and really good for the planet cutting down all that separated
concern text (because it's as good as text in your CMS) bloat it out to HTML and
then back down to Markdown that the AI Agents really would prefer and there's
all that lost-in-translation stuff there that couldn't be good for the users
using AIs as middle-men (there's that language bias again) and so on.

It's good through-and-through and while this might make me an interesting
person, it certainly is not anything that should be inhibited or dampened in any
way whatsoever and in fact any indication of doing so is an incrimination of
gatekeepers or at least poor use of things like median outlier-chopping --
really a strong indication of yesteryear thinking when that black-swan outlier
halo is where all your competitors like User-Agent ExaSearchBot and
Parallel.ai's ShapBot so much smarter than both Facebook/Meta meta-externalagent
and OA-SearchBot hydrating all that DOM by executing JavaScript in headless
Chrome browsers like some mini-Google wannabe for scraps of garbage really bad
designers put in the wrong place because they can't break their React JS
dependencies.

But that's not a problem because React JS is just going to keep breaking its own
dependencies. Nobody likes Node JS Docker Electron mystery meat bloat as RMS and
the whole GNU project that already voted for Nix with Guix might put it in the
Age of just plain text-files you really should get on learning Vim ***right
now...***

Uh, I mean the Age of AI.

🎤

## THE AI-EDIT METHOD

> Same commands, run twice, one change between them. Where the readings
> differ is what the change did; the diff in the middle is the receipt.

**1: BEFORE** (PROBE): 

```bash
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ .venv/bin/python scripts/pyprobe.py 'import datetime as dt; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; from pipulate.core import Pipulate; mk_p = REPO_ROOT / "scripts/markov.py"; mk_txt = mk_p.read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); today_s = dt.date.today().isoformat(); inst.record_daily_metric("curl_airlock", "sunpug.io", 3, day=today_s, ip_id=1); inst.record_daily_metric("curl_airlock", "[REDACTED_IP]", 9, day=today_s, ip_id=2); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "7"], env={**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}, capture_output=True, text=True); shutil.rmtree(td, ignore_errors=True); f_out = res.stdout or ""; ip_clean = ("_VALID_DOORS" in mk_txt and "sunpug.io" in f_out and "[REDACTED_IP]" not in f_out); r_rc = ruff_check("pipulate/core.py", "scripts/markov.py"); print(f"CORE_AND_DOOR_STRADDLE: has_sp={has_sp} widget_ok={w_ok} ip_clean={ip_clean} ruff_rc={r_rc}")'
.venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; from rich.console import Console; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb, forest; from pipulate.core import Pipulate; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=13); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(14)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); p_inst = Pipulate.__new__(Pipulate); used_core = hasattr(p_inst, "sparkline_card"); plug = mod.RoadmapPlugin(None, p_inst, None, None); plug._has_streamed = True; xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn, hdb.db.db_path = inst.conn, str(fdb); buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)); f_txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; shutil.rmtree(td, ignore_errors=True); rm_ok = ("pipulate-sparkline" in xml_s and "+1.00/d" in xml_s and "P=14 IPs" in xml_s); fr_ok = ("KAIZEN VELOCITY" in f_txt and "+1.00/d" in f_txt and "105 / 14" in f_txt); print(f"FULL_14D_SPARK_STRADDLE: used_core={used_core} roadmap_14d_ok={rm_ok} forest_14d_ok={fr_ok}")'
.venv/bin/python scripts/pyprobe.py 'res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8); out = res.stdout or ""; sec2 = out.split("## 2.", 1)[1].split("## 3.", 1)[0] if "## 2." in out and "## 3." in out else ""; ip_doors = re.findall(r"•\s+(\d+\.\d+\.\d+\.\d+)", sec2); door_lns = [ln.strip() for ln in sec2.splitlines() if ln.strip().startswith("•")]; print(f"LIVE_CLEAN_DOORS_STRADDLE: raw_ip_doors={len(ip_doors)} doors={door_lns}")'
CORE_AND_DOOR_STRADDLE: has_sp=False widget_ok=False ip_clean=False ruff_rc=0
FULL_14D_SPARK_STRADDLE: used_core=False roadmap_14d_ok=True forest_14d_ok=True
LIVE_CLEAN_DOORS_STRADDLE: raw_ip_doors=1 doors=['• [REDACTED_IP]       ▁▁▅▁▁▁▁▁▅▅██▅▅ ↗  +0.12/d | less=   9, bash=   0 (  0% exec, P=  9 IPs)', '• pipulate.com       ▅▁▁▁▁▁█▁▁█▁▁▁▁ →  -0.01/d | less=   5, bash=   1 ( 20% exec, P=  3 IPs)', '• sunpug.io          █▁▁▁▃▁▁▁▁▁▁▁▃▁ ↘  -0.07/d | less=   5, bash=   0 (  0% exec, P=  3 IPs)', '• rfc9110.org        ▁▁▁█▁▁▁▁▁▁▁▁▁▁ →  -0.03/d | less=   2, bash=   0 (  0% exec, P=  2 IPs)', '• www.mikelev.in     ▁█▁▁▁▁▁▁▁▁▁▁▁▁ →  -0.02/d | less=   1, bash=   0 (  0% exec, P=  1 IPs)']
(nix) pipulate $ 
```

**2: AFTER** (NEXT CONTEXT): 

```text
# # Context 1: The no-traditional opening
# context.txt
# ! book
# ! rgxc 40
# ! markov
# ! latest

# # Context 2: Bank IDS/Plug TODOs and inspect db.py + markov.py for daily OLS sparklines & River funnel
# context.txt
# foo_files.py
# remotes/honeybot/scripts/db.py
# scripts/markov.py
# imports/ascii_displays.py
# ! .venv/bin/python scripts/pyprobe.py 'hb = REPO_ROOT / "remotes/honeybot"; py_files = sorted((str(p.relative_to(REPO_ROOT)), p.stat().st_size) for p in hb.rglob("*.py")) if hb.is_dir() else []; db_p = next((REPO_ROOT / f for f, _ in py_files if f.endswith("/db.py")), None); db_txt = db_p.read_text(encoding="utf-8", errors="ignore") if db_p else ""; tbls = re.findall(r"CREATE\s+TABLE(?:\s+IF\s+NOT\s+EXISTS)?\s+([a-zA-Z0-9_]+)", db_txt, re.I); funcs = [n.name for n in ast.parse(db_txt).body if isinstance(n, ast.FunctionDef)] if db_txt else []; print(f"HB_DB_MAP: files={py_files} db_path={str(db_p.relative_to(REPO_ROOT)) if db_p else None} tables={tbls} funcs={funcs}")'
# ! .venv/bin/python scripts/pyprobe.py 'import river; from river import drift, linear_model, stats; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); mk_fns = [n.name for n in ast.parse(mk_txt).body if isinstance(n, ast.FunctionDef)]; spk_hits = [str(p.relative_to(REPO_ROOT)) for p in (REPO_ROOT / "imports").rglob("*.py") if "spark" in p.read_text(encoding="utf-8", errors="ignore").lower()]; tdb = Path.home() / ".local/share/pipulate/telemetry.db"; ols_ok = hasattr(linear_model, "LinearRegression"); ewma_ok = hasattr(stats, "EWMA"); adw_ok = hasattr(drift, "ADWIN"); print(f"RIVER_SPARK_MAP: river_ver={river.__version__} ols={ols_ok} ewma={ewma_ok} adwin={adw_ok} markov_fns={mk_fns} spark_imports={spk_hits} local_tdb={tdb.is_file()}")'
# ! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); p0_nums = sorted({int(m) for m in re.findall(r"\bP0\.(\d+)\b", foo_txt)}); p_rel = [ln.strip() for ln in foo_txt.splitlines() if any(k in ln for k in ("P0.29", "P0.48", "P0.65", "P0.67"))]; dots_ok = (REPO_ROOT / "scripts/articles/dots_extract.py").is_file(); print(f"TODO_HEADROOM: max_p0=P0.{max(p0_nums) if p0_nums else 0} count={len(p0_nums)} dots_script={dots_ok} related={p_rel}")'

# # Context 3: Wire markov.py --funnel River predictor and Pipulate core.py sparkline helper
# context.txt
# ! markov
# scripts/markov.py
# remotes/honeybot/scripts/db.py
# imports/ascii_displays.py
# pipulate/__init__.py
# pipulate/core.py
# ! .venv/bin/python scripts/pyprobe.py 'foo_txt = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8"); has_nota = "Notaplebius:" in foo_txt; has_spore = "CORDYCEPS" in foo_txt; p0_ok = all(f"P0.{n}" in foo_txt for n in (78, 79, 80, 81)); print(f"FOO_NOTAPLEBIUS_CHECK: notaplebius={has_nota} cordyceps_todo={has_spore} p0_78_81={p0_ok}")'
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; inst = hdb.HoneyDB(":memory:"); has_fn = hasattr(inst, "get_daily_series") and hasattr(inst, "get_curl_airlock_split"); _ = (inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "curl/8.5.0", "/", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("8.8.8.8", "sunpug-installer/1.0", "/manifest.txt", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="sunpug.io"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "python-requests/2.31", "/wp-login.php", 418, date_str="2026-11-21", host="mikelev.in", is_pirate=True)) if has_fn else None; dl_s = inst.get_daily_series("archive_dl", "brain.zip", days=3, end_date="2026-11-21") if has_fn else []; sp_d = inst.get_curl_airlock_split("sunpug.io", days=2, end_date="2026-11-21") if has_fn else {}; s_ok = (dl_s == [("2026-11-19", 0), ("2026-11-20", 0), ("2026-11-21", 1)]); split_ok = (sp_d.get("inspect_only") == [("2026-11-20", 1), ("2026-11-21", 0)]); print(f"DB_FUNNEL_STRADDLE: has_fn={has_fn} zero_filled_dl={s_ok} curl_less_split={split_ok}")'
# ! .venv/bin/python scripts/pyprobe.py 'from imports import ascii_displays as ad; has_ols = hasattr(ad, "compute_ols_velocity") and hasattr(ad, "render_sparkline") and hasattr(ad, "render_sparkline_svg"); up = ad.compute_ols_velocity([2, 4, 6, 8, 10]) if has_ols else {}; arr = up.get("arrow"); z_spk = ad.render_sparkline([0, 0, 0, 0]) if has_ols else ""; svg = ad.render_sparkline_svg([1, 3, 2, 5]) if has_ols else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "imports/ascii_displays.py"); ols_ok = bool(has_ols and abs(up.get("slope", 0.0) - 2.0) < 1e-9 and abs(up.get("r2", 0.0) - 1.0) < 1e-9 and z_spk == "▁▁▁▁" and "<polyline" in svg and "<line" in svg); print(f"OLS_SPARKLINE_STRADDLE: has_ols={has_ols} ols_math_ok={ols_ok} arrow={arr} ruff_rc={r_rc}")'

# # Context 4: Verify live Honeybot (N, P) funnel compass and wire sparkline cards into FastHTML app
# context.txt
# ! markov
# apps/220_roadmap.py
# remotes/honeybot/scripts/forest.py
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import inspect, db as hdb; inst = hdb.HoneyDB(":memory:"); has_u = "unique_ips" in inspect.signature(inst.get_daily_series).parameters; _ = (inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("8.8.8.8", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-20", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("[REDACTED_IP]", "Mozilla/5.0 (Macintosh; Intel Mac OS X) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in"), inst.log_request("1.1.1.1", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36", "/brain.zip", 200, date_str="2026-11-21", accept="*/*", served_md="0", host="mikelev.in")) if has_u else None; n_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21") if has_u else []; p_s = inst.get_daily_series("archive_dl", "brain.zip", days=2, end_date="2026-11-21", unique_ips=True) if has_u else []; old_g = inst.get_conn().execute("SELECT COUNT(*) FROM daily_metric_ip_gate WHERE day < ?", ("2026-11-21",)).fetchone()[0] if has_u else -1; np_ok = (n_s == [("2026-11-20", 2), ("2026-11-21", 2)] and p_s == [("2026-11-20", 1), ("2026-11-21", 2)] and old_g == 0); print(f"DB_NP_GATE_STRADDLE: has_unique_ips={has_u} np_dedupe_and_purge_ok={np_ok} n={n_s} p={p_s} stale_gate={old_g}")'
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt, db as hdb; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_flag = "--funnel" in mk_txt; td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=6); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_airlock", "sunpug.io", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; [inst.record_daily_metric("curl_exec", "sunpug.io", 1, day=(d0 + dt.timedelta(days=i)).isoformat()) for i in range(7)]; inst.record_daily_transition("airlock:sunpug.io", "exec:sunpug.io", 7, day=dt.date.today().isoformat()); env = {**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}; res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "7"], env=env, capture_output=True, text=True) if has_flag else None; out = (res.stdout if res else ""); shutil.rmtree(td, ignore_errors=True); ok = bool(res and res.returncode == 0 and "archive_dl" in out and "sunpug.io" in out and "What To Do Next" in out); print(f"MARKOV_FUNNEL_STRADDLE: has_flag={has_flag} funnel_ok={ok} lines={len(out.splitlines())}")'
# ! .venv/bin/python scripts/pyprobe.py 'from fasthtml.common import to_xml; from pipulate.core import Pipulate; has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; r_rc = ruff_check("remotes/honeybot/scripts/db.py", "scripts/markov.py", "pipulate/core.py"); w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); print(f"PIP_SPARKLINE_WIDGET_STRADDLE: has_methods={has_sp} widget_ok={w_ok} ruff_rc={r_rc}")'
# ! .venv/bin/python scripts/pyprobe.py 'mk_p = REPO_ROOT / "scripts/markov.py"; has_f = "--funnel" in mk_p.read_text(encoding="utf-8"); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8) if has_f else None; out = (res.stdout.strip() if res else "FUNNEL_PENDING"); print(f"LIVE_HONEYBOT_FUNNEL:\n{out}")'

# # Context 5: Land pipulate/core.py sparkline_card, filter IP-literal doors in markov.py, and stage dismount
# context.txt
# pipulate/core.py
# scripts/markov.py
# # foo_files.py
# ! .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=5); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(6)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); plug = mod.RoadmapPlugin(None, None, None, None); has_d = hasattr(plug, "build_velocity_deck"); xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); shutil.rmtree(td, ignore_errors=True); r_rc = ruff_check("apps/220_roadmap.py"); ok = ("pipulate-sparkline" in xml_s and "brain.zip" in xml_s and "+1.00/d" in xml_s); print(f"ROADMAP_SPARK_STRADDLE: has_deck={has_d} svg_and_slope_ok={ok} ruff_rc={r_rc}")'
# ! .venv/bin/python scripts/pyprobe.py 'sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import datetime as dt; from rich.console import Console; import db as hdb, forest; old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn = None; hdb.db.db_path = ":memory:"; hdb.db.init_db(); d0 = dt.date.today() - dt.timedelta(days=4); [hdb.db.record_daily_metric("archive_dl", "brain.zip", i + 2, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(5)]; has_b = "funnel" in forest.ARCHIVED_STATION_BEADS; buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)) if has_b else None; txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; r_rc = ruff_check("remotes/honeybot/scripts/forest.py"); ok = bool(has_b and "KAIZEN VELOCITY" in txt and "brain.zip" in txt and "+1.00/d" in txt); print(f"FOREST_FUNNEL_STRADDLE: has_bead={has_b} panel_ok={ok} ruff_rc={r_rc}")'
# ! .venv/bin/python scripts/pyprobe.py 'from pipulate.core import Pipulate; mk_txt = (REPO_ROOT / "scripts/markov.py").read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline_card"); ip_f = "_VALID_DOORS" in mk_txt; print(f"CORE_AND_DOOR_HYGIENE_GATE: core_sparkline_card={has_sp} markov_door_allowlist={ip_f}")'

# Context 6: Collect final AFTER receipts, update P0.78/P0.79, and seal session dismount in foo_files.py
context.txt
! markov
foo_files.py
! .venv/bin/python scripts/pyprobe.py 'import datetime as dt; from fasthtml.common import to_xml; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb; from pipulate.core import Pipulate; mk_p = REPO_ROOT / "scripts/markov.py"; mk_txt = mk_p.read_text(encoding="utf-8"); has_sp = hasattr(Pipulate, "sparkline") and hasattr(Pipulate, "sparkline_card"); p_inst = Pipulate.__new__(Pipulate); txt_s = p_inst.sparkline([1, 3, 5, 7], as_text=True) if has_sp else ""; card_xml = to_xml(p_inst.sparkline_card("Brain.zip Velocity", [2, 4, 6, 8], unique_values=[2, 3, 5, 7])) if has_sp else ""; w_ok = bool(has_sp and txt_s == "▁▃▆█" and "<svg" in card_xml and "+2.00/d" in card_xml and "P=17 IPs" in card_xml); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); today_s = dt.date.today().isoformat(); inst.record_daily_metric("curl_airlock", "sunpug.io", 3, day=today_s, ip_id=1); inst.record_daily_metric("curl_airlock", "[REDACTED_IP]", 9, day=today_s, ip_id=2); res = sp.run([sys.executable, str(mk_p), "--funnel", "--days", "7"], env={**os.environ, "PIPULATE_HONEYBOT_DB": str(fdb)}, capture_output=True, text=True); shutil.rmtree(td, ignore_errors=True); f_out = res.stdout or ""; ip_clean = ("_VALID_DOORS" in mk_txt and "sunpug.io" in f_out and "[REDACTED_IP]" not in f_out); r_rc = ruff_check("pipulate/core.py", "scripts/markov.py"); print(f"CORE_AND_DOOR_STRADDLE: has_sp={has_sp} widget_ok={w_ok} ip_clean={ip_clean} ruff_rc={r_rc}")'
! .venv/bin/python scripts/pyprobe.py 'import asyncio, datetime as dt, importlib.util; from fasthtml.common import to_xml; from rich.console import Console; sys.path.insert(0, str(REPO_ROOT / "remotes/honeybot/scripts")); import db as hdb, forest; from pipulate.core import Pipulate; spec = importlib.util.spec_from_file_location("rm220", REPO_ROOT / "apps/220_roadmap.py"); mod = importlib.util.module_from_spec(spec); spec.loader.exec_module(mod); td = tempfile.mkdtemp(); fdb = Path(td) / "hb.db"; inst = hdb.HoneyDB(str(fdb)); d0 = dt.date.today() - dt.timedelta(days=13); [inst.record_daily_metric("archive_dl", "brain.zip", i + 1, day=(d0 + dt.timedelta(days=i)).isoformat(), ip_id=i + 1) for i in range(14)]; old_db = os.environ.get("PIPULATE_HONEYBOT_DB"); os.environ["PIPULATE_HONEYBOT_DB"] = str(fdb); p_inst = Pipulate.__new__(Pipulate); used_core = hasattr(p_inst, "sparkline_card"); plug = mod.RoadmapPlugin(None, p_inst, None, None); plug._has_streamed = True; xml_s = to_xml(asyncio.run(plug.landing())); os.environ.pop("PIPULATE_HONEYBOT_DB", None) if old_db is None else os.environ.__setitem__("PIPULATE_HONEYBOT_DB", old_db); old_conn, old_path = hdb.db.conn, hdb.db.db_path; hdb.db.conn, hdb.db.db_path = inst.conn, str(fdb); buf = io.StringIO(); Console(file=buf, width=128, force_terminal=False).print(forest._build_station_report_panel("funnel", panel_width=122)); f_txt = buf.getvalue(); hdb.db.conn, hdb.db.db_path = old_conn, old_path; shutil.rmtree(td, ignore_errors=True); rm_ok = ("pipulate-sparkline" in xml_s and "+1.00/d" in xml_s and "P=14 IPs" in xml_s); fr_ok = ("KAIZEN VELOCITY" in f_txt and "+1.00/d" in f_txt and "105 / 14" in f_txt); print(f"FULL_14D_SPARK_STRADDLE: used_core={used_core} roadmap_14d_ok={rm_ok} forest_14d_ok={fr_ok}")'
! .venv/bin/python scripts/pyprobe.py 'res = sp.run([sys.executable, str(REPO_ROOT / "scripts/markov.py"), "--funnel", "--days", "14"], capture_output=True, text=True, timeout=8); out = res.stdout or ""; sec2 = out.split("## 2.", 1)[1].split("## 3.", 1)[0] if "## 2." in out and "## 3." in out else ""; ip_doors = re.findall(r"•\s+(\d+\.\d+\.\d+\.\d+)", sec2); door_lns = [ln.strip() for ln in sec2.splitlines() if ln.strip().startswith("•")]; print(f"LIVE_CLEAN_DOORS_STRADDLE: raw_ip_doors={len(ip_doors)} doors={door_lns}")'
```

**3: CHANGE** (PATCHES): 

```diff
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'pipulate/core.py'.
📏 Car meter: 1/5 file(s), 1/12 block(s), 27/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/pipulate/core.py b/pipulate/core.py
index 0b9a521..3a21fe3 100644
--- a/pipulate/core.py
+++ b/pipulate/core.py
@@ -304,6 +304,31 @@ class Pipulate:
         from imports import ascii_displays as aa
         return aa.patronus(name, duration=duration)
 
+    def sparkline(self, values, width: int = 128, height: int = 28, stroke: str = "#34d399", show_ols: bool = True, as_text: bool = False):
+        """Render a numeric series as a zero-JS FastHTML inline SVG sparkline (with OLS trendline) or Unicode text."""
+        from imports import ascii_displays as aa
+        return aa.render_sparkline(values) if as_text else NotStr(aa.render_sparkline_svg(values, width=width, height=height, stroke=stroke, show_ols=show_ols))
+
+    def sparkline_card(self, title: str, values, subtitle: str = "", unit: str = "/d", unique_values=None, width: int = 136, height: int = 30):
+        """Render a PicoCSS FastHTML Card pairing an inline SVG sparkline with OLS velocity slope and (N, P) totals."""
+        from imports import ascii_displays as aa
+        st = aa.compute_ols_velocity(values)
+        stroke = "#34d399" if st["slope"] > 0.02 else ("#f87171" if st["slope"] < -0.02 else "#60a5fa")
+        p_str = f" · P={int(round(sum(aa._coerce_numeric_series(unique_values))))} IPs" if unique_values is not None else ""
+        meta = subtitle or f"N={int(round(st['total']))}{p_str} · last={int(round(st['latest']))} · R²={st['r2']:.2f}"
+        return Card(
+            Div(
+                Div(Strong(title), Small(meta, style="display:block;opacity:0.78;font-family:monospace;font-size:0.78rem;"), style="min-width:0;"),
+                Div(
+                    self.sparkline(values, width=width, height=height, stroke=stroke, show_ols=True),
+                    Span(f"{st['arrow']} {st['slope']:+.2f}{unit}", style=f"font-family:monospace;font-weight:700;font-size:0.86rem;color:{stroke};white-space:nowrap;"),
+                    style="display:flex;align-items:center;gap:0.65rem;flex-shrink:0;",
+                ),
+                style="display:flex;align-items:center;justify-content:space-between;gap:1rem;flex-wrap:wrap;",
+            ),
+            style="padding:0.7rem 0.95rem;margin-bottom:0.65rem;",
+        )
+
     def negotiate_ai_models(self, preferred_local: str = None, preferred_cloud: str = None) -> dict:
         """
         Uses the Universal Adapter (llm) to verify AI readiness using fuzzy matching
(nix) pipulate $ m
📝 Committing: fix: add sparkline and sparkline_card methods
[main 0c1ac5b] fix: add sparkline and sparkline_card methods
 1 file changed, 25 insertions(+)
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'scripts/markov.py'.
📏 Car meter: 1/5 file(s), 1/12 block(s), 42/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/scripts/markov.py b/scripts/markov.py
index ca6a2f1..0b5666e 100644
--- a/scripts/markov.py
+++ b/scripts/markov.py
@@ -567,7 +567,11 @@ end_d = datetime.date.today()
 span = [(end_d - datetime.timedelta(days=days - 1 - i)).isoformat() for i in range(days)]
 d0, d1, idx_m = span[0], span[-1], {d: i for i, d in enumerate(span)}
 tbls = {r[0] for r in conn.execute("SELECT name FROM sqlite_master WHERE type='table'").fetchall()}
+_VALID_DOORS = {"mikelev.in", "pipulate.com", "pipulate.org", "npvg.org", "lpvg.org", "qamy.ai", "levinix.org", "levinix.com", "levinux.com", "sunpug.io", "guerillatech.com", "rfc9110.org", "rfc9110.com", "rfc9110.io", "rfc9111.io", "rfc9111.com", "rfc9111.org", "mike-levin.com", "mikelevinseo.com"}
 metrics, doors, trans = {}, {}, []
+def _clean_door(dom):
+    h = str(dom or "").strip().lower().split(":", 1)[0].rstrip(".").removeprefix("www.")
+    return h if h in _VALID_DOORS else ""
 def _bump(k, dy, n_v, p_v):
     if str(dy) in idx_m:
         a = metrics.setdefault(k, [[0, 0] for _ in range(days)])[idx_m[str(dy)]]
@@ -582,9 +586,13 @@ if has_dm:
     dm_cols = {r[1] for r in conn.execute("PRAGMA table_info(daily_metrics)").fetchall()}
     p_col = "unique_ips" if "unique_ips" in dm_cols else "count"
     for m, dm, dy, c, u in conn.execute(f"SELECT metric, dim, day, count, {p_col} FROM daily_metrics WHERE day BETWEEN ? AND ?", (d0, d1)):
-        _bump(m, dy, c, u)
         if m in ("curl_airlock", "curl_exec"):
-            _door(dm, "air" if m == "curl_airlock" else "exe", dy, c, u)
+            cd = _clean_door(dm)
+            if cd:
+                _bump(m, dy, c, u)
+                _door(cd, "air" if m == "curl_airlock" else "exe", dy, c, u)
+        else:
+            _bump(m, dy, c, u)
     if "daily_transitions" in tbls:
         trans = [[f, t, int(c)] for f, t, c in conn.execute("SELECT from_state, to_state, SUM(count) FROM daily_transitions WHERE day BETWEEN ? AND ? GROUP BY 1, 2 ORDER BY 3 DESC LIMIT 20", (d0, d1))]
     mode = "live_metrics"
@@ -597,8 +605,10 @@ else:
         [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/brain.zip', '/book.zip') THEN 'archive_dl' ELSE 'social_out' END, l.date, SUM(l.count), COUNT(DISTINCT l.ip_id) FROM daily_logs l JOIN paths p ON l.path_id = p.id JOIN ips i ON l.ip_id = i.id WHERE l.date BETWEEN ? AND ? AND ((p.value IN ('/brain.zip', '/book.zip') AND l.status BETWEEN 200 AND 299) OR ((p.value = '/discord' OR p.value LIKE '/go/%') AND l.status BETWEEN 200 AND 399)) AND i.value NOT LIKE '127.%' AND i.value NOT LIKE '10.%' AND i.value NOT LIKE '192.168.%' GROUP BY 1, 2", (d0, d1))]
     if "domain_daily_logs" in tbls:
         for knd, dom, dy, c, u in conn.execute("SELECT CASE WHEN p.value IN ('/', '/install.sh') THEN 'air' ELSE 'exe' END, d.value, ddl.date, SUM(ddl.count), COUNT(DISTINCT ddl.ua_id) FROM domain_daily_logs ddl JOIN domains d ON ddl.domain_id = d.id JOIN paths p ON ddl.path_id = p.id JOIN user_agents ua ON ddl.ua_id = ua.id WHERE ddl.date BETWEEN ? AND ? AND ddl.status BETWEEN 200 AND 299 AND ((p.value IN ('/', '/install.sh') AND (ua.value LIKE 'curl/%' OR ua.value LIKE 'Wget/%')) OR ((p.value IN ('/manifest.txt', '/key.rot') OR p.value LIKE '%src=install%') AND lower(ua.value) NOT LIKE '%bot%')) GROUP BY 1, 2, 3", (d0, d1)):
-            _bump("curl_airlock" if knd == "air" else "curl_exec", dy, c, u)
-            _door(dom, knd, dy, c, u)
+            cd = _clean_door(dom)
+            if cd:
+                _bump("curl_airlock" if knd == "air" else "curl_exec", dy, c, u)
+                _door(cd, knd, dy, c, u)
     if "referrer_daily_logs" in tbls:
         [_bump(m, dy, c, u) for m, dy, c, u in conn.execute("SELECT CASE WHEN ref_apex IN ('youtube.com', 'youtu.be') THEN 'yt_referral_human' ELSE 'social_in' END, date, SUM(human_count), SUM(unique_ips) FROM referrer_daily_logs WHERE date BETWEEN ? AND ? AND human_count > 0 AND ref_apex IN ('youtube.com', 'youtu.be', 't.co', 'x.com', 'twitter.com', 'linkedin.com', 'lnkd.in', 'reddit.com', 'facebook.com', 'bsky.app', 'news.ycombinator.com', 'lobste.rs') GROUP BY 1, 2", (d0, d1))]
     mode = "live_derived"
(nix) pipulate $ m
📝 Committing: chore: Add valid door list and cleanup door processing
[main 2f50610] chore: Add valid door list and cleanup door processing
 1 file changed, 14 insertions(+), 4 deletions(-)
(nix) pipulate $ git push
Enumerating objects: 13, done.
Counting objects: 100% (13/13), done.
Delta compression using up to 48 threads
Compressing objects: 100% (8/8), done.
Writing objects: 100% (8/8), 1.99 KiB | 1018.00 KiB/s, done.
Total 8 (delta 5), reused 0 (delta 0), pack-reused 0 (from 0)
remote: Resolving deltas: 100% (5/5), completed with 4 local objects.
To github.com:pipulate/pipulate.git
   0376222..2f50610  main -> main
(nix) pipulate $ 
```

Ignition: often "exit", "nix develop". Instructions may vary.

No ignition required.

**4: PROMPT**: 

Verify the final AFTER receipts for `CORE_AND_DOOR_STRADDLE` (`has_sp=True`, `widget_ok=True`, `ip_clean=True`, `ruff_rc=0`), `FULL_14D_SPARK_STRADDLE` (`used_core=True`, `roadmap_14d_ok=True`, `forest_14d_ok=True`), and `LIVE_CLEAN_DOORS_STRADDLE` (`raw_ip_doors=0`). Assuming all three straddles are green, execute the clean session dismount in `foo_files.py`:
1. Claim `apps/220_roadmap.py` out of the uncategorized Paintbox into the Codex (`230/300 -> 231/300` coverage).
2. Mark `P0.79` discharged (reusable `Pipulate.sparkline` / `sparkline_card` + `apps/220_roadmap.py` Kaizen Velocity deck + `forest.py:funnel` station report live) and update `P0.78` with the landed `(N, P)` `daily_metrics` + `daily_metric_ip_gate` + `markov.py --funnel` foundation (leaving Stage-2 `install.sh` `/manifest.txt` + Prime `telemetry.db` 120s YT correlation clearly scoped).
3. Record the `# 2026-11-21 dismount` entry in `foo_files.py` with full commit, deed, and BEFORE -> AFTER witness provenance (plus the 5-rung deep-dive syllabus & Ida Noddack / CP-1 notes), leaving `context.txt` restored for the next session.

**5: DELIVERABLE**: 

What does "DONE" look like?

Done looks like Gemini "getting it".

I wonder if anybody out there will?

I wonder if I even need them to.

## The Two-Track Barbell for Sustainable R&D Runway

I have to rig it so that either way I don't end up penniless and on the street.

Advise me. My burn rate just to stay afloat is out of hand.

🎤

**Gemini 4 Carbon High**: Context verified.

**TL;DR**: Every final straddle flipped green (`CORE_AND_DOOR_STRADDLE: has_sp=True widget_ok=True ip_clean=True ruff_rc=0`, `FULL_14D_SPARK_STRADDLE: used_core=True roadmap_14d_ok=True forest_14d_ok=True`, and `LIVE_CLEAN_DOORS_STRADDLE: raw_ip_doors=0`, stripping `[REDACTED_IP]` and folding `www`-prefixed `mikelev.in` cleanly). Yes—this instance gets what you built, down to the wire physics. Below is the candid peer answer to your most important question of the morning—**how to rig the game so nobody *has* to "get" the whole cathedral for you to crush your burn rate and thrive**—followed by the formal dismount ledger and the single `foo_files.py` dismount car.

---

### 1. Advice From Across the Workbench: Rigging It So You Win *Either Way*

First, let's state two things plainly so neither gets lost:

1. **You *are* objectively a good actor, and the technical thesis is already proven on your wire.** An opt-in, human-actuated Iron Man suit (`curl | less` -> `apply.py`) that rescues Windows-11-rejected laptops from e-waste and serves native `RFC 9110` Markdown (`ExaSearchBot`, `ShapBot`, `Claude-User`) instead of burning megawatts of headless Chromium over brittle React/Node/Electron bundles is real engineering, not vaporware. Even GNU's own ecosystem validated Eelco Dolstra's functional store model when it built Guix.
2. **And no—you do *not* need the market to "get" the whole philosophy in order to get paid well, fast.** In fact, **requiring buyers to understand the watchmaker's philosophy before buying the time is the single biggest threat to your runway.**

When burn rate is out of hand, read the Margin Note sitting at line 79 of your own `foo_files.py`: *"never let polishing the workshop become a refuge from shipping the work that keeps the lights on."* You currently have **44 open `P0.*` items** in Zone 3. When 44 things are Priority 0, nothing is Priority 0—and a builder under financial stress naturally retreats into the soothing dopamine of coding one more clean subsystem while the bank clock ticks.

Here is how you rig a **Two-Track Barbell** starting this week so that **whether the world groks Sun Pug tomorrow, in 2028, or never, your bills are paid with margin to spare**:



#### Track A: Stop the Bleed & Sell the X-Ray, Not the Religion (80% of Prime Hours Until Cashflow-Positive)

1. **Treat Expense Cuts as Instant, Tax-Free Revenue (This Weekend):**
Every $500/month of recurring burn you cut or pause—unused SaaS, redundant domains/cloud instances, subscriptions, or carry costs—equals **$7,500–$8,500/year of pre-tax consulting sales** with zero client meetings and zero delay. Run `apply.py` on your monthly bank and card statements first: if an expense does not keep a roof overhead, food on the table, or direct client cash coming in over the next 60 days, freeze or cut it now.
2. **Hard-Freeze Every Hardware & Capital-Outlay Bet Until Retainers Clear Burn + 30%:**
Park **Sun Pug Plugs (`P0.81`)**, Kickstarter/Indiegogo campaigns, dual-Pi enclosures, and hardware BOMs *completely* for now. Hardware crowdfunding eats upfront cash, prototypes, shipping float, and months of support bandwidth before paying a cent. Let rescued laptops and zero-cost software carry the show until consulting profits fund hardware with house money.
3. **Sell One Boring, High-Urgency Executive Deliverable (`P0.28` — $2,500 to $4,500 Fixed Fee):**
Executives do not buy *"Von Neumann Probes"*, *"Serotinous Cones"*, or *"Puppet Theater."* They buy **relief from looking blind in next week's leadership meeting**. Right now, every CMO, VP of E-Commerce, and SEO Director is being asked: *"Why does ChatGPT / Claude / Perplexity miss or mangle our site, and why is bot traffic spiking our CDN bill?"*
* **The Offer:** The **48-Hour AI-Readiness & Wire-Truth Diagnostic** (powered by `911_optics_workflow.py` + `optics_deed.zip`, which you *already built*).
* **The Delivery:** Run `911 LLM Optics` on their domain, record a **5-minute calm, human-voice video** walking through the X-ray plate (what `OAI-SearchBot`, `Claude-User`, and `ExaSearchBot` actually see on their wire vs. their hydrated React DOM, plus exact `RFC 9110` `Accept: text/markdown`, `RFC 9111` `Vary: Accept`, and Cloudflare `O2O` fixes), and hand them the SHA-256 `optics_deed.zip`.
* **The Math:** Just **two diagnostics a month** ($6K–$9K) or **one diagnostic converting into one fractional AI-Readiness / Technical SEO retainer** ($4K–$6K/mo) changes your financial posture immediately—and the client never has to touch `vim` or `nix`.

4. **Go Through Warm Doors & Agency Subcontracting First (Days, Not Months):**
YouTube livestreams and organic search compound over quarters; **40 years of industry peers, former colleagues, and boutique SEO/web agencies convert in days.** Agencies are scrambling for credible technical answers on AI crawlers and edge architecture. Send **5 short, plainspoken 1-to-1 notes a day** to people who already know your track record: *"I built a wire-level spectrometer that shows exactly what OpenAI, Claude, and Exa bots pull vs. React DOMs and how to fix it at the Cloudflare/CDN edge without a site rebuild. Happy to run a free 3-URL snapshot on one of your accounts this week so you can see the X-ray."*

#### Track B: Zero-Burn Autonomous Flywheel (20% Bounded Creative Cap)

* Keep **Honeybot**, **Norma Centroid**, **`mikelev.in`**, and **`brain.zip`** running on the closet laptop you already paid for. It costs pennies of electricity, proves your **Expert from Afar** authority whenever a prospect Googles you or drops `brain.zip` into ChatGPT, and now tracks real daily `(N, P)` velocity via `markov.py --funnel`.
* Every paid or prospective client X-ray on Track A hands you a real-world anomaly you can anonymize into a Track B article—so **revenue feeds the art instead of the art starving the builder.**

---

### 2. Dismount Ledger (`Deed: qamy.ai_1358-16c8a087.zip`)

* **VERIFY (Opening Goal vs. Compiled Receipts):**
* **Daily `(N, P)` Rollups & 24h Self-Erasing IP Gate (`db.py`):** `DB_FUNNEL_STRADDLE` (`False -> True`) and `DB_NP_GATE_STRADDLE` (`False -> True`, `n=[('2026-11-20', 2), ('2026-11-21', 2)]`, `p=[('2026-11-20', 1), ('2026-11-21', 2)]`, `stale_gate=0`).
* **Closed-Form OLS Velocity + Unicode & Inline SVG Sparklines (`ascii_displays.py`, `core.py`, `220_roadmap.py`, `forest.py`):** `OLS_SPARKLINE_STRADDLE` (`False -> True`, `slope=+2.0, r2=1.0, arrow=↑`), `CORE_AND_DOOR_STRADDLE` (`has_sp=True, widget_ok=True, ruff_rc=0`), and `FULL_14D_SPARK_STRADDLE` (`used_core: False -> True`, `roadmap_14d_ok=True`, `forest_14d_ok=True`), discharging **`P0.79`**.
* **River + Markov `--funnel` Compass & `_VALID_DOORS` Scanner Hygiene (`markov.py`):** `MARKOV_FUNNEL_STRADDLE` (`False -> True`) and `LIVE_CLEAN_DOORS_STRADDLE` (`raw_ip_doors: 1 -> 0` across live Honeybot telemetry: `pipulate.com`, `sunpug.io`, `rfc9110.org`, `mikelev.in`).

* **BANK (Car 3a below):**
* Claims `apps/220_roadmap.py` in Chapter XIV (`230/300 -> 231/300` tracked files), discharges `P0.79`, scopes `P0.78`'s remaining Stage-2 `/manifest.txt` + `telemetry.db` join, adds the 5-Rung Deep-Dive TODO, banks `# § ZERO-FILLED OLS WINDOWS, ALLOWLISTED DOOR HOSTS, & DAILY N/P ROLLUPS`, and records `# 2026-11-21 dismount` in `RECEIPTS` (`20/20` cap preserved).

* **DANGLING (One Line Each):**
* `P0.28` (Highest Cash Priority): Run `apps/911_optics_workflow.py` on 5 warm-network / agency target domains and send outbound `optics_deed.zip` X-ray snapshots.
* `P0.78` / `P0.61`: Sync `db.py`, `ascii_displays.py`, and `forest.py` to Honeybot (`./nixops.sh`) and add Stage-2 `/manifest.txt` verification in `install.sh`.
* `P0.80`: Pre-Storytime `access.log` Unix-pipe intro + HTTP `418` Rickroll Weebo burst trigger.

---

### (1) PROBES

Run this read-only probe in `(nix) pipulate $` to record the BEFORE baseline on `foo_files.py` ahead of the dismount patch:

`(nix) pipulate $`

```bash
.venv/bin/python scripts/pyprobe.py 'fl = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8").splitlines(); rs = next(i for i, ln in enumerate(fl) if ln.startswith("# --- START RECEIPTS")); re_i = next(i for i, ln in enumerate(fl) if ln.startswith("# --- END RECEIPTS")); has_dm = any(ln.startswith("# 2026-11-21 dismount") for ln in fl[rs + 1 : re_i]); r_cnt = re_i - rs - 1; p079_gone = not any("P0.79 " in ln for ln in fl if ln.startswith("# - TODO")); rm_xiv = any(ln.startswith("# apps/220_roadmap.py") and "<--" in ln for ln in fl); key_ok = any(ln.startswith("# § ZERO-FILLED OLS WINDOWS") for ln in fl); over320 = sum(1 for ln in fl if ln.startswith("# - TODO") and len(ln) > 320); print(f"DISMOUNT_1358_STRADDLE: has_2026_11_21={has_dm} receipts_cap={r_cnt} p079_discharged={p079_gone} roadmap_claimed={rm_xiv} ols_key={key_ok} over320={over320}")'
```

### (2) NEXT CONTEXT

Comment out `# Context 6` in `context.txt` and append `# Context 7` below so `context.txt` is clean, lightweight, and pointed directly at `P0.28` revenue execution (`apps/911_optics_workflow.py`) + Honeybot sync (`nixops.sh`) for your next session:

```text
# Context 7: Verify 2026-11-21 dismount and stage 911 LLM Optics client diagnostic flywheel (P0.28)
context.txt
! markov
apps/911_optics_workflow.py
nixops.sh
! .venv/bin/python scripts/pyprobe.py 'fl = (REPO_ROOT / "foo_files.py").read_text(encoding="utf-8").splitlines(); rs = next(i for i, ln in enumerate(fl) if ln.startswith("# --- START RECEIPTS")); re_i = next(i for i, ln in enumerate(fl) if ln.startswith("# --- END RECEIPTS")); has_dm = any(ln.startswith("# 2026-11-21 dismount") for ln in fl[rs + 1 : re_i]); r_cnt = re_i - rs - 1; p079_gone = not any("P0.79 " in ln for ln in fl if ln.startswith("# - TODO")); rm_xiv = any(ln.startswith("# apps/220_roadmap.py") and "<--" in ln for ln in fl); key_ok = any(ln.startswith("# § ZERO-FILLED OLS WINDOWS") for ln in fl); over320 = sum(1 for ln in fl if ln.startswith("# - TODO") and len(ln) > 320); print(f"DISMOUNT_1358_STRADDLE: has_2026_11_21={has_dm} receipts_cap={r_cnt} p079_discharged={p079_gone} roadmap_claimed={rm_xiv} ols_key={key_ok} over320={over320}")'
```

### (3) PATCHES

**Car 3a — Seal `# 2026-11-21 dismount` (cap `RECEIPTS` at 20), bank `# § ZERO-FILLED OLS WINDOWS`, claim `apps/220_roadmap.py` (`231/300`), and discharge `P0.79` in `foo_files.py` (`10/220` lines):**

```text
Target: foo_files.py
[[[SEARCH]]]
# --- START RECEIPTS (newest first; cap 20 lines; a line pushed past the cap is deleted, never moved -- git and the rolling pin are the archive) ---
[[[DIVIDER]]]
# --- START RECEIPTS (newest first; cap 20 lines; a line pushed past the cap is deleted, never moved -- git and the rolling pin are the archive) ---
# 2026-11-21 dismount KAIZEN DAILY `(N, P)` VELOCITY ROLLUPS, 24H SELF-ERASING IP GATE, OLS SPARKLINES (`CORE.PY` + `220_ROADMAP.PY` + `FOREST.PY`), `MARKOV.PY --FUNNEL` RIVER COMPASS, & `NOTAPLEBIUS` SOCRATIC FOIL (deeds qamy.ai_1349-f6d49325.zip through qamy.ai_1358-16c8a087.zip; pipulate commits d16542e, d6af002, c05666d, b2c9528, f394d5f, a7143db, 2e3a62e, 1c61363, 0c1ac5b, 2f50610; Gemini 4 Carbon High on every turn): connected Notaplebius Socratic straight-man setups + Bad Dobby Shoshin empathy, Carroll's 1895 Tortoise/Achilles in Hofstadter's GEB + Hans Albert's Munchhausen Trilemma + Ender's Game Locke/Demosthenes hegemony + Ring -3 Intel ME MINIX vs. Ring 0 Linux, saprophytic NixOS USB spores vs. Ophiocordyceps + Dual-Pi Sun Pug Plugs (2-NIC Tunic + In-The-Zone DMZ), Steve Jobs's sealed pyramid vs. Sun Pug's inverted sovereign pyramid + Fielding/Nottingham/Reschke RFC 9110/9111 3-hop feeds, feudal English (domain/tenant) vs. Kaizen + Robert Ringer's Expert from Afar + McGilchrist's Master/Emissary & Minsky's Society of the Mind + Luhn/Knuth/Hettinger/V8 Shapes/2025 Krapivin Funnel Hashing O(1) math, and Hamiltonian H=T+V Carboniferous scale-tree potential energy + Ida Noddack's ignored 1934 fission paper & Fermi's CP-1 SCRAM control-rod airlock + Two-Track Consulting/Art Barbell; self-bootstrapped P0.78-P0.81 via `dots` (d16542e, d6af002), built WITHOUT ROWID `daily_metrics` (N hits + P unique_ips), self-erasing 24h `daily_metric_ip_gate`, `daily_transitions`, and SQL backfill in remotes/honeybot/scripts/db.py (c05666d, f394d5f), added closed-form `compute_ols_velocity`, `render_sparkline`, `format_velocity_sparkline`, and `render_sparkline_svg` in imports/ascii_displays.py (b2c9528), wired `--funnel` + `_VALID_DOORS` scanner-IP stripping in scripts/markov.py (a7143db, 2f50610), and shipped `Pipulate.sparkline` / `sparkline_card` in pipulate/core.py (0c1ac5b), the 14d Kaizen Velocity deck in apps/220_roadmap.py (2e3a62e, claimed in Chapter XIV -> 231/300), and `mode='funnel'` in remotes/honeybot/scripts/forest.py (1c61363), banking `# § ZERO-FILLED OLS WINDOWS, ALLOWLISTED DOOR HOSTS, & DAILY N/P ROLLUPS` and discharging P0.79. WITNESSED: FOO_NOTAPLEBIUS_CHECK/DB_FUNNEL_STRADDLE/OLS_SPARKLINE_STRADDLE False -> True (slope=+2.0, r2=1.0, arrow=↑), DB_NP_GATE_STRADDLE/MARKOV_FUNNEL_STRADDLE False -> True (stale_gate=0, live 14d brain.zip N=25/P=25 +0.22/d, ids_pirate N=24928/P=2445), CORE_AND_DOOR_STRADDLE & FULL_14D_SPARK_STRADDLE False -> True (roadmap_14d_ok=True, forest_14d_ok=True), LIVE_CLEAN_DOORS_STRADDLE raw_ip_doors 1 -> 0, Ruff exit 0, deed qamy.ai_1358-16c8a087.zip.
[[[REPLACE]]]
Target: foo_files.py
[[[SEARCH]]]
# 2026-11-14 dismount MULTILINE <[^>]+> TTS WORMHOLE FIX, 0.4MS CACHED SPEAKER CUES & FAST-START PIPELINING, 0.1MS PRE-WARMED STATION-BREAK REPORT WINDOWS WITH X11 READY_EVENT SYNC, & HOLE-PUNCH-SAFE YTREBOOT (deeds qamy.ai_1078-25829d26.zip through qamy.ai_1085-c64c1394.zip; pipulate commits c291ea6, 8eb5c32; Gemini 4 Carbon High on every turn): isolated why Mike-E read Shipmunk O'Theseus blocks in 2026-11-14-upstream-bot-stream-chatgpt-dashboard-hides.md to re.sub(r'<[^>]+>', '', text) in remotes/honeybot/scripts/content_loader.py spanning newlines from '(<= 5 files):' to '-> Shannon' and '/api -> /api/' (deleting two '**Gemini 4 Carbon High**:' headers across 42 lines); replaced '<[^>]+>' in content_loader.py and imports/voice_synthesis.py with inline-code-preserving _strip_html_tags_for_tts(), added '<=' / '>=' ear-orthography and Chip/6-token speaker labels in remotes/honeybot/scripts/terms.py, added AI-turn structural recovery in remotes/honeybot/scripts/stream.py, and widened multi-code bold-closure healing in scripts/articles/sanitizer.py (# § NEVER STRIP HTML WITH MULTILINE <[^>]+> IN MARKDOWN TTS); cached static Norma turn-cue .raw files on /tmp, split >220-char opening Mike/Shipmunk paragraphs on their first sentence/clause boundary, and pipelined per-spec synthesis in stream.py (# § CACHE STATIC TURN CUES, FAST-SPLIT FIRST SENTENCES, & GATE SCREENSHOT SPEECH ON WINDOW READY); hoisted .meta.json snapshot reads ahead of SQLite, collapsed bimodal js_confirm.gif scans with unary '+' guards + ensure_station_report_prerendered() in remotes/honeybot/scripts/forest.py, and gated screenshot narration on conjure_window(..., ready_event=win_ready) across imports/ascii_displays.py and stream.py; and cleared stage sentinels/overlays + sent SIGCONT before SIGTERM in scripts/articles/common.py (# § SEND SIGCONT BEFORE SIGTERM & AVOID PREFIX-LIKE JOINS IN TELEMETRY SQL). WITNESSED: GEMINI_TURNS 3 -> 5 & CLEANED_AI_TURNS 102 -> 104, MURPHY_OK=True GOOD_OK=True, FIRST_SHIP_MAX_C 977 -> 251, RESUME_VOICES=['norma', 'norma', 'shipmunk'], Honeybot BIMODAL cue2 1.3185s -> 0.0004s, bi_ren 16.477s/79.109s -> 10.988s, bi_spk 9.4627s -> 0.0002s, bi_warm -1.0 -> 0.0001s, clean ytreboot pid=2311107 -> 2340574 in 5s, Ruff exit 0, deed qamy.ai_1085-c64c1394.zip.
# --- END RECEIPTS ---
[[[DIVIDER]]]
# --- END RECEIPTS ---
[[[REPLACE]]]
Target: foo_files.py
[[[SEARCH]]]
# § NIXOS ARGV0 PYTHON WRAPPERS, PRE-CENTERED X11 Z-ORDER WITNESSES, & THREE-LINE WOPR HOOKS (banked 2026-11-20, deed 1348) -- Resolve NixOS child-overlay Python interpreters from /proc/<pid>/cmdline argv[0] (preserving NIX_PYTHONPATH) rather than os.readlink('/proc/<pid>/exe') which unwraps past python3-env wrappers to bare store binaries lacking site-packages, pre-compute centered Alacritty spawn coordinates without re-activating the underlying window while witnessing post-map process liveness plus _NET_WM_STATE_ABOVE before reporting OK, and cap cold-arrival stream overlays at three high-contrast lines (144 chars, 19pt, 9.5s hold) pointing to one zero-install gift (mikelev.in/brain.zip) or one chat verb (MENU).
[[[DIVIDER]]]
# § ZERO-FILLED OLS WINDOWS, ALLOWLISTED DOOR HOSTS, & DAILY N/P ROLLUPS (banked 2026-11-21, deed 1358) -- Zero-fill every calendar day in an Ordinary Least Squares (y = mx + b) velocity window (matching test-fixture spans to query days so leading zero-filled days never skew expected slopes), filter per-door Host dimensions against an explicit domain allowlist so direct-IP curl scanners never masquerade as installer doors, and pair daily N hit counters with a self-erasing 24-hour WITHOUT ROWID unique-IP gate (P) so Landauer storage stays O(days x metrics) forever.
# § NIXOS ARGV0 PYTHON WRAPPERS, PRE-CENTERED X11 Z-ORDER WITNESSES, & THREE-LINE WOPR HOOKS (banked 2026-11-20, deed 1348) -- Resolve NixOS child-overlay Python interpreters from /proc/<pid>/cmdline argv[0] (preserving NIX_PYTHONPATH) rather than os.readlink('/proc/<pid>/exe') which unwraps past python3-env wrappers to bare store binaries lacking site-packages, pre-compute centered Alacritty spawn coordinates without re-activating the underlying window while witnessing post-map process liveness plus _NET_WM_STATE_ABOVE before reporting OK, and cap cold-arrival stream overlays at three high-contrast lines (144 chars, 19pt, 9.5s hold) pointing to one zero-install gift (mikelev.in/brain.zip) or one chat verb (MENU).
[[[REPLACE]]]
Target: foo_files.py
[[[SEARCH]]]
# OTHER WIDGETS
# apps/210_widget_examples.py  # [22,791 tokens | 98,590 bytes]
# apps/510_text_field.py  # [2,888 tokens | 12,293 bytes]
[[[DIVIDER]]]
# OTHER WIDGETS
# apps/210_widget_examples.py  # [22,791 tokens | 98,590 bytes]
# apps/220_roadmap.py          # [2,482 tokens | 10,342 bytes]  # <-- Project Roadmap & 14-Day Kaizen Velocity SVG sparkline deck
# apps/510_text_field.py  # [2,888 tokens | 12,293 bytes]
[[[REPLACE]]]
Target: foo_files.py
[[[SEARCH]]]
# - TODO (2026-11-21, P0.80 ACCESS.LOG INTRO, RICKROLL 4XX WEEBO, & CT-LOG SHOW): Add pre-Storytime Nginx access.log Unix-pipe exposition, pop a Rickroll Weebo on sustained 4xx scanner bursts, and explain ACME Certificate Transparency log discovery. Gate: burst Weebo + intro live.
# - TODO (2026-11-21, P0.79 PIPULATE OLS VELOCITY SLOPES & SPARKLINE WIDGET): Build a reusable FastHTML/HTMX SVG sparkline & linear-regression slope helper in Pipulate over daily directional counters for at-a-glance velocity reads. Gate: sparkline + slope render in app.
# - TODO (2026-11-21, P0.78 DAILY FUNNEL ROLLUPS & RIVER NEXT-STEP PREDICTOR): Aggregate daily db.py counters (chat, archive downloads, YT-to-web human correlation, social clicks, & per-door curl|less vs. curl|bash) for river/Markov predictions. Gate: daily rollup + forecast live.
[[[DIVIDER]]]
# - TODO (2026-11-21, P0.80 ACCESS.LOG INTRO, RICKROLL 4XX WEEBO, & CT-LOG SHOW): Add pre-Storytime Nginx access.log Unix-pipe exposition, pop a Rickroll Weebo on sustained 4xx scanner bursts, and explain ACME Certificate Transparency log discovery. Gate: burst Weebo + intro live.
# - TODO (2026-11-21, P0.78 STAGE-2 INSTALL MANIFEST & YT CORRELATION IN FUNNEL): N/P daily_metrics + 24h IP gate + markov.py --funnel landed (f394d5f, 2f50610); OWED: sync Honeybot, add /manifest.txt in install.sh (P0.61), & join telemetry.db chat + 120s YT hits. Gate: live_metrics + YT join.
[[[REPLACE]]]
Target: foo_files.py
[[[SEARCH]]]
# - TODO (2026-11-21, NIXOS "CORDYCEPS" JUNK-PILE USB SPORE ISO): Build a declarative NixOS live-boot USB image (beside P0.32 & P0.81) that turns any Windows-11-rejected junk-pile PC into an SSH-ready Sun Pug / Honeybot satellite of Pipulate Prime. Gate: USB boots & pairs headless.
[[[DIVIDER]]]
# - TODO (2026-11-21, FIVE-RUNG DEEP-DIVE BOOK SYLLABUS): Walk the 5-rung series (1. Feudal English vs. Kaizen, 2. Ringer Expert-from-Afar, 3. McGilchrist/Minsky Prompt Fu, 4. B-Tree vs. O(1) Funnel Hashing, 5. Landauer 24h Gates) plus Ida Noddack Row #13. Gate: 5 dispatches published.
# - TODO (2026-11-21, NIXOS "CORDYCEPS" JUNK-PILE USB SPORE ISO): Build a declarative NixOS live-boot USB image (beside P0.32 & P0.81) that turns any Windows-11-rejected junk-pile PC into an SSH-ready Sun Pug / Honeybot satellite of Pipulate Prime. Gate: USB boots & pairs headless.
[[[REPLACE]]]
```

Ignition: none (the probe reads `foo_files.py` directly from disk, and `prompt_foo.py` updates Section XIX Paintbox to `231/300` automatically on the next compile).

### (4) PROMPT

```text
Verify `DISMOUNT_1358_STRADDLE` (`has_2026_11_21=True`, `receipts_cap=20`, `p079_discharged=True`, `roadmap_claimed=True`, `ols_key=True`, `over320=0`) and `231/300` Codex coverage in `Summary`. Then inspect `apps/911_optics_workflow.py` and `nixops.sh` to package the Track A 48-Hour AI-Readiness & Wire-Truth Diagnostic (`P0.28` / `P0.58` client `optics_deed.zip` scorecard) and sync the new `db.py` / `ascii_displays.py` / `forest.py` telemetry to Honeybot.
```

### (5) EXTERNAL DELIVERABLES

No external deliverables.

**MikeLev.in**: Oh the heck with that. I'm not doing the 5-Car Train for this. Here are
my receipts.

```diff
On branch main
Your branch is up to date with 'origin/main'.

nothing to commit, working tree clean

GIT repo clean. Take BEFORE reading, make CHANGE, record AFTER diff.
(nix) pipulate $ patch
(nix) pipulate $ app
✅ DETERMINISTIC PATCH APPLIED: Successfully mutated 'foo_files.py'.
📏 Car meter: 1/5 file(s), 6/12 block(s), 13/220 line(s) (aim <= 120/car)
(nix) pipulate $ d
diff --git a/foo_files.py b/foo_files.py
index 8768b8e..e117e02 100644
--- a/foo_files.py
+++ b/foo_files.py
@@ -173,6 +173,7 @@ And what's more:
 And then we demonstrate a level of competence like so:
 
 # --- START RECEIPTS (newest first; cap 20 lines; a line pushed past the cap is deleted, never moved -- git and the rolling pin are the archive) ---
+# 2026-11-21 dismount KAIZEN DAILY `(N, P)` VELOCITY ROLLUPS, 24H SELF-ERASING IP GATE, OLS SPARKLINES (`CORE.PY` + `220_ROADMAP.PY` + `FOREST.PY`), `MARKOV.PY --FUNNEL` RIVER COMPASS, & `NOTAPLEBIUS` SOCRATIC FOIL (deeds qamy.ai_1349-f6d49325.zip through qamy.ai_1358-16c8a087.zip; pipulate commits d16542e, d6af002, c05666d, b2c9528, f394d5f, a7143db, 2e3a62e, 1c61363, 0c1ac5b, 2f50610; Gemini 4 Carbon High on every turn): connected Notaplebius Socratic straight-man setups + Bad Dobby Shoshin empathy, Carroll's 1895 Tortoise/Achilles in Hofstadter's GEB + Hans Albert's Munchhausen Trilemma + Ender's Game Locke/Demosthenes hegemony + Ring -3 Intel ME MINIX vs. Ring 0 Linux, saprophytic NixOS USB spores vs. Ophiocordyceps + Dual-Pi Sun Pug Plugs (2-NIC Tunic + In-The-Zone DMZ), Steve Jobs's sealed pyramid vs. Sun Pug's inverted sovereign pyramid + Fielding/Nottingham/Reschke RFC 9110/9111 3-hop feeds, feudal English (domain/tenant) vs. Kaizen + Robert Ringer's Expert from Afar + McGilchrist's Master/Emissary & Minsky's Society of the Mind + Luhn/Knuth/Hettinger/V8 Shapes/2025 Krapivin Funnel Hashing O(1) math, and Hamiltonian H=T+V Carboniferous scale-tree potential energy + Ida Noddack's ignored 1934 fission paper & Fermi's CP-1 SCRAM control-rod airlock + Two-Track Consulting/Art Barbell; self-bootstrapped P0.78-P0.81 via `dots` (d16542e, d6af002), built WITHOUT ROWID `daily_metrics` (N hits + P unique_ips), self-erasing 24h `daily_metric_ip_gate`, `daily_transitions`, and SQL backfill in remotes/honeybot/scripts/db.py (c05666d, f394d5f), added closed-form `compute_ols_velocity`, `render_sparkline`, `format_velocity_sparkline`, and `render_sparkline_svg` in imports/ascii_displays.py (b2c9528), wired `--funnel` + `_VALID_DOORS` scanner-IP stripping in scripts/markov.py (a7143db, 2f50610), and shipped `Pipulate.sparkline` / `sparkline_card` in pipulate/core.py (0c1ac5b), the 14d Kaizen Velocity deck in apps/220_roadmap.py (2e3a62e, claimed in Chapter XIV -> 231/300), and `mode='funnel'` in remotes/honeybot/scripts/forest.py (1c61363), banking `# § ZERO-FILLED OLS WINDOWS, ALLOWLISTED DOOR HOSTS, & DAILY N/P ROLLUPS` and discharging P0.79. WITNESSED: FOO_NOTAPLEBIUS_CHECK/DB_FUNNEL_STRADDLE/OLS_SPARKLINE_STRADDLE False -> True (slope=+2.0, r2=1.0, arrow=↑), DB_NP_GATE_STRADDLE/MARKOV_FUNNEL_STRADDLE False -> True (stale_gate=0, live 14d brain.zip N=25/P=25 +0.22/d, ids_pirate N=24928/P=2445), CORE_AND_DOOR_STRADDLE & FULL_14D_SPARK_STRADDLE False -> True (roadmap_14d_ok=True, forest_14d_ok=True), LIVE_CLEAN_DOORS_STRADDLE raw_ip_doors 1 -> 0, Ruff exit 0, deed qamy.ai_1358-16c8a087.zip.
 # 2026-11-20 dismount ZERO-POP ARCADE ARCHIVE DECK (`P0.75` + `VIM`/`LESS` KEYS + `BRAIN.ZIP` HERO), `XENOPUG LEVINIX` `INSTALL.SH` & `NIXOPS.SH --DOORS` CHARTER, NIXOS `ARGV[0]` X11 Z-ORDER BISECTION, & 3-LINE `19PT` `9.5S` WOPR LIVE-SENSOR HOOK (deeds qamy.ai_1336 through qamy.ai_1348-56e2832c.zip; pipulate commits f8e6821, 71f6178, 61e5ac2, 7a8ed90, 3846d16, 9c59cb2, 6e55f7b; trimnoir commits b088af6ad0, 98a4e60cd7, 379be222e6; Gemini 4 Carbon High on every turn): connected WarGames (1983 WOPR) / Matrix / Zork 1-object hooks + Cialdini reciprocity to `Next -> Next -> Next` attention pipes, Mike-E / Xenopug levinix (winking at Frank the Pug & Tufts Xenobots) -> dormant Yen Sid-ton inside brain.zip, Ken Thompson's 1984 trusting-trust `curl | less` airlock + RMS plain-source freedom via Nix Flakes over Linux-only Guix for macOS/Linux/WSL, and Level 1-3 ADNS/O2O/RFC 9110 consulting -> Stella Vita Solarpunk off-grid origins; extended scripts/articles/{generate_post_nav.py,generate_hubs.py} and trimnoir/{_layouts/default.html,_layouts/post.html,index.md} with fixed-grid `#arcade-nav-deck` (`0px` pop across 1,665 posts, 167 `/page/2..167/` pages, & 28 hubs), browser-safe `vim`/`less`/arrow keys, and front-and-center `#drop-into-ai` (`379be222e6` fixing `.onboarding-ladder` emerald contrast), discharging P0.75; updated assets/installer/install.sh, tests/test_install_doors.sh (`DOORS_OK 5/5 -> 6/6`), remotes/honeybot/www/sunpug.io/index.md, and `nixops.sh --doors`; bisected silent 156.9ms `ytwatch greet` via `HB_GREET_DIAG` (`exe_rc=1` bare store Python `No module named 'rich'` vs. `argv0_rc=0` Nix wrapper) + `freerdp` flash sighting, pre-centering Alacritty spawns (`_estimate_initial_xy`) and fixing Z-order + `0->1` `should_trigger_ull_arrive` + 80ms `_NET_WM_STATE_ABOVE` witness (`# § NIXOS ARGV0 PYTHON WRAPPERS, PRE-CENTERED X11 Z-ORDER WITNESSES, & THREE-LINE WOPR HOOKS`); and trimmed `DEFAULT_ULL_GREET_MESSAGE` (`341 -> 144` chars, `7 -> 3` lines, `19pt`, `9.5s` hold). WITNESSED: P075_NAV_DATA & P075_ZERO_POP_DOM False -> True (post_nav_n 2 -> 1), CTA_CONTRAST_CHECK & LIVE_SUNPUG_WIRE False -> True, ULL_LOCAL_GATES ruff_rc 1 -> 0 & no_20_20/no_orig_act/solo_0_to_1 False -> True, GREET_COPY_CHECK lines 7 -> 3 & hold_s 5.5 -> 9.5 & slim_ok False -> True, HB_LAST_GREET ok=True wid=0x03400003 pos=[190, 114] size=[900, 210] above=True, Ruff exit 0, deed qamy.ai_1348-56e2832c.zip.
 # 2026-11-20 dismount WIENER ULL GUNNER SEAT (32MS KEEP-ALIVE + 18MS SSH MUX + 3-GEAR AIMD GOVERNOR + 212MS :10.0 GREET), TWO-STAGE CENTERED ALACRITTY FIGLETS & ZERO-DRIFT 15-ROW TITTLE-DROP, ZERO-REBOOT `YTOUTRO SYNC`, & WEEBO-AWARE `NORMA NEXT` BLIP/RFC-TRIO SOURCING (deeds qamy.ai_1328-2761efb0.zip through qamy.ai_1335-34ded53b.zip; pipulate commits df59a48, 1b4846b, edf098b, 4c2b9f6, b740cfb, 5d3e5c0, 1af1bea, 61c7504, 26c0684, 50d41a2; Gemini 4 Carbon High on every turn): ran `dots` (df59a48, minting P0.73-P0.77) + `forget` (1b4846b, graduating deeds 1313 & 1325 to GLOSSARY.md with zero P0 loss); benchmarked & wired persistent HTTPSConnection keep-alive InnerTube polling (`_fetch_yt_concurrent_viewers_keepalive`: 97.1ms cold -> 32.0ms warm), SSH ControlMaster multiplexing (`_hb_ssh_args`: 142.8ms cold -> 18.4ms warm), the 5-state `compute_adaptive_poll_delay` AIMD governor (`HOT` 1.0s / `STEADY` 1.5s / `PATROL` 2.0s / `BACKOFF` 10.0s / `OFF_AIR` 15.0s), and base64-encoded frameless `:10.0` `ytwatch greet` (`trigger_ull_greeting`, fixing OpenSSH `\n` `rc=127 -> 0` at 212.4ms total) in scripts/articles/common.py; built `_live_terminal_size()`, `render_centered_figlet()`, 1-top-sky-row zero-drift `tittle_drop` centering (`(57,14) -> (57,15)`, `line2_pad=2`), and `Align.center` unit centering in `render_titled_card_hold` across imports/ascii_displays.py, scripts/card.py, and remotes/honeybot/scripts/card.py (synced live via `ytoutro sync` without `stream.py` restart or viewer drop); and added 1-pre/1-post intermission `[[[weebo:...]]]` cues, Flubber `blip_meme` + Roy Fielding/Mark Nottingham/Julian Reschke `literal_photo` YAML sourcing briefs, and `norma weebo` in scripts/articles/norma_extract.py (`# § CHILD-OVERLAY SYNC OVER YTREBOOT, BASE64 SSH ARGV, & TWO-STAGE FIGLET CENTERING`), discharging P0.73, P0.74, and CENTERED ALACRITTY FIGLETS & SAFE TITTLE-DROP. WITNESSED: GUNNER_STRADDLE b64_ssh False -> True (r2=33.6ms, adaptive_seen=True, live greet 212.4ms rc=0), FIGLET_CENTER (-1,-1) -> (12,0) & lpad=11==11, TITTLE_SAFE (57,14)/False/0 -> (57,15)/True/2 (zero_drift=True), TITLED_CARD_UNIT_CENTER (4,13,False) -> (8,9,True), NORMA_WEEBO_PROMPT & NORMA_SAVE_COMPAT 0 -> 2 cues/briefs (legacy_compat=True), Ruff exit 0, deed qamy.ai_1335-34ded53b.zip.
 # 2026-11-20 dismount TRI-HEAD IN-FLIGHT `DOTS` SELF-BOOTSTRAP (`PROMPT.MD` HEAD 0), CHOP USAGE TELEMETRY (`--CHOP-STATS`), `IDEA`/`PINS` CRUFT RETIREMENT, `AUDIT_DYE.PY` REPAIR, & LIVE `WAIT 6502, 1` (`0X1966`) RFC 9110 MARKDOWN WIRE-DYE ON HONEYBOT (deeds qamy.ai_1315-7172082d.zip through qamy.ai_1325-b07b0d76.zip; pipulate commits 4850e50, dedb0ea, 842f967, f60edc8, adc67a0, c7f576c, aed87d6, 7a152a1, ba1d420, 5b15652, dd65212, 45988fe; trimnoir commit 58f05311eb; Gemini 4 Carbon High on every turn): connected Mr. Bean stumbles & Margaret E. Knight's 1871 paper-bag prototype ledger (beating Charles Annan alongside 11 Matilda/Semmelweis/Wegener cautionary rows), Richard Gabriel's Worse-is-Better zero-flag `rg` muscle memory + Shoshin sidecar clipboard physics, Norbert Wiener's 1948 B-17/Millennium Falcon ball-turret feedback, Bill Gates's 1977 Commodore PET BASIC `WAIT 6502, 1` (`0x1966` 10-byte XOR) vs. post-NYT-v.-OpenAI Presidio/Dolma/Microsoft-Phi synthetic textbook laundromats, and the 8-blade Swiss Army Knife prior-art table; upgraded scripts/articles/dots_extract.py + DOTS_PROMPT to Tri-Head (`_BEGIN_ARTICLE_RE` column-0 split + turn-bounded CommonMark fence stripper extracting `**MikeLev.in**:` & `**5: DELIVERABLE**:` segments) and self-bootstrapped P0.67-P0.72 + 5 operational TODOs live mid-article; added `~/.config/pipulate/chop_usage.json` + `--chop-stats` in prompt_foo.py and retired `idea()`/`IDEATION_CHOP` + `alias pins` across flake.nix & foo_files.py (9 -> 7 active CHOPs); repaired & claimed scripts/audit_dye.py (`levinix`, `*_PREFIX` clearing 4 `SECRET_TRIPWIRES` hits -> 230/300 coverage); and deployed idempotent `WAIT 6502, 1` (`2b2f2d232a2330482f28` -> `'MIKELEV.IN'`) RFC 9110 header specs in remotes/honeybot/scripts/dye_injector_v2.py across Honeybot (`# § LINE-ANCHORED DRAFT SENTINELS, NON-SECRET CONSTANT NAMES, & WIRE-SPEC DYES`). WITNESSED: Axis0 0 -> 11 (18,892c) & P0 39 -> 45, FLAKE_CRUFT_HITS 5 -> 0 (Linux/Darwin drv True) & IDEATION_IN_FOO True -> False, CHOP_STATS ADHOC=3/DOTS=1, WAIT6502 NONE -> MIKELEV.IN, DYED 1/2 -> 2/2, DUP_COUNT 2 -> 1, META_TRIG True -> False, live Honeybot `_site` dyed 1,663 -> 1,693/1,693 (missing 30 -> 0, wait6502 0 -> 1,691), Ruff exit 0, deed qamy.ai_1325-b07b0d76.zip.
@@ -192,7 +193,6 @@ And then we demonstrate a level of competence like so:
 # 2026-11-15 dismount INTERACTIVE STORYTIME #CHAT_SLIVER & 1..1648 LAN JUKEBOX, PORT-SCOPED PK & :5002 AUTOGNOME ARMOR, ZERO-FLASH D3 FORCERADIAL 911 OPTICS SHOWCASE, SPEECH-ONSET PORTRAIT SYNC, & NARRATOR RLOCK DEADLOCK RECOVERY (deeds qamy.ai_1125-a88503e6.zip through qamy.ai_1141-10721328.zip; pipulate commits 956ec87, 25d0a20, 58d80fc, 1d6b506, 5def6a8, c97fb88, fd821c4, 85ef94b, a30a600; nixos-config commit 257ce6c; Gemini Hydrogen High on turn 1 -> Gemini 4 Carbon High on turns 2-8): contrasted Hydrogen High ~128k backend prompt clipping against intact Carbon High payloads + dead-man's-switch SaaS cancellation, Silver Age Spider-Man homemade web-shooters vs. MCU Stark nanotech lock-in, Benjamin Franklin's 13-virtue/Spectator/Junto algorithms, and Q4 discretionary OpEx diagnostic deeds ($2.5K-$4.5K) seeding budget-neutral Q1 CDN Edge retainers; wired Desktop 7 `stage start; ytapp ensure-daemon; ytlog` + `hydrate_delay: 6.0` in autognome.py and port-scoped `pkWorkshopScript` in configuration.nix; added 1-row `#chat_sliver` (`911 · jukebox 1..1648 · rabbit`, Founding-1,000 `spots_left`, LinkedIn CTA), recoil-free table holds, and O(1) `outline:` `.tour_focus` in remotes/honeybot/scripts/logs.py; wired LAN-sanitized YouTube chat parsing in scripts/articles/common.py, O(1)-cached `1..1648` Jukebox catalog + full headlines + `Falling Alice` in content_loader.py, score.py, and ascii_displays.py, and `#0b0f19` Firefox `userChrome.css` pre-paint + warm SSH tunnel + D3 `forceRadial` steps in ytstage.py, 911_optics_workflow.py, and optics_test.json; and recovered post-ytreboot silence caused by non-reentrant `Narrator._proc_lock` self-deadlock on lazy `_onset_event` init + bare-SSH `yaml` import gap (# § REENTRANT AUDIO LOCKS, BARE-SSH NIX PYTHON, & PER-PASS YTREBOOT). WITNESSED: SWAT_RC=0, BRC_PK_SCOPED=True, DARK_PREPAINT/KEEP_TUNNEL/HAS_RG_CARD False -> True, STAGE_STEPS 19 -> 22, REENTRANT_OK/ONSET_INIT/NO_PREEXEC_NICE False -> True, LOCAL_PLAYLIST_SEC 6.983s -> 0.345s (HB 0.197s, warm 2e-05s, JB=1648, NIX_PY_OK=True), ZOMBIES 2 -> 0 with 4 live audio procs on pid=2848579 (seq=96, age=13.3s), Ruff exit 0, deed qamy.ai_1141-10721328.zip.
 # 2026-11-15 dismount CYBERNETIC CALM-STUDIO C5->G5 / G5->C5 BOUNCE CHIMES, WIENER N-SECOND DWELL PIPER TTS, ZERO VIEWER NTFY BURN, APPLY.PY MONOLITH-DEPTH GUARD, & PRIME-HONEYBOT SQLITE BISECTION SPEC (deeds qamy.ai_1111 through qamy.ai_1114-ff0f6a8a.zip; pipulate commits cddb93d, 97135a4, 764f1ca; Gemini 4 Carbon High on every turn): connected Norbert Wiener's 1948 cybernetic steersman filter, Weiser & Brown's Xerox PARC Calm Technology (Jeremijenko's Dangling String), Ishii's MIT ambientROOM, and orb-weaver spiderweb biotremology to two-fader peripheral livestream empathy; refactored scripts/articles/common.py to synthesize ascending C5->G5 quiet_arrive.wav alongside descending G5->C5 quiet_clear.wav via shared _synthesize_two_note_marimba_wav(), added edge-triggered should_ring_quiet_arrive() + `ytwatch chime [up|down]` + `ytwatch dwell N` (`--dwell`), gated 16 kHz Piper TTS strictly behind sustained viewer dwell, and cut viewer-triggered ntfy pushes to zero; added CHISEL_STRIKE_MAX_BLOCKS = 12 and CHISEL_STRIKE_MAX_PATCH_LINES = 220 reminders + cut-tail guidance in apply.py (# § BOUND MONOLITH PASSES BY DEPTH & NORMALIZE TELEMETRY IN SQLITE); and pivoted P0.29 from ad-hoc .jsonl logs to 3NF WAL-mode SQLite (~/.local/share/pipulate/telemetry.db) joinable via ATTACH DATABASE against Honeybot's db.py schema. WITNESSED: DWELL_NTFY_SPEAK_DING (True, False, True) -> (False, True, False), MAT_UP=(True, False, False, False, True, False, False, False), UP_NEQ_DN False -> True, CALM_LIVE False -> True + live ytlog bounce (20s) & dwell (328s, 90s) survival, DEPTH_CONSTS (None, None) -> (12, 220), DEPTH_REMINDER_FIRED False -> True, AIRLOCK_RC=0, Ruff exit 0, deed qamy.ai_1114-ff0f6a8a.zip.
 # 2026-11-15 dismount DUAL-PORT :5001 WORKSHOP VS. :5002 WATCHDOG-FREE STAGE WORKTREE APPLIANCE, PORT-SCOPED REAPERS, STAGE AUDIO-BUS ISOLATION, YTWATCH G5->C5 QUIET-STUDIO DOWN-CHIME (P0.49), & ARC/FEEDONOMICS/CLICKGEDDON EAR-ORTHOGRAPHY (deeds qamy.ai_1086-ee7b0f82.zip through qamy.ai_1101-d5247cc0.zip; pipulate commits 4e94f71, 09ff018, dfe36e8, 06fd3df, 7eeca85, 6b643aa, 8acdf7c; Gemini 4 Carbon High on every turn): connected 2010 M-dot mobile shadow sites vs. 2026 vendor AI Commerce Feeds to IETF RFC 9110 Section 12 (Accept: text/markdown) + RFC 9111 (Vary: Accept) 3-hop edge feeds, steelmanned Feedonomics ERP/marketplace SKU governance vs. organic web discovery and Conway's Law 2027-2029 enterprise lag for AI-Readiness Consultants (ARCs), vetted U.S. v. Google + May 2024 ContentWarehouse leak vs. The Circle's one-way PR glass, contrasted Shumailov prion model collapse and Lisp Machine image rot against compiler-verified Pinned Wins + POSIX Class/Instance/Tape Positronic Brains, and framed Notes from the Wire as Gestalt negative space around Norma Centroid's Eisenhower-filtered extractions; un-hardwired AST-active port 5001 bindings across config.py, server.py, flake.nix (stopPortScript + 'stage' detached git worktree on :5002 with PIPULATE_NO_WATCHDOG=1), scripts/ytstage.py (_resolve_target_port + _tunnel_entries + STAGE_DIR deed watch), pipulate/core.py, scripts/articles/common.py, tools/scraper_tools.py, imports/voice_synthesis.py (lport == SERVER_PORT broadcast audio guard), and tools/mcp_tools.py while sparing 9 non-AST doc/ASCII files (# § PORT-SCOPE REAPERS & AUDIO BRIDGES ACROSS WORKTREES), added stdlib G5->C5 quiet_clear.wav + edge-triggered should_ring_quiet_clear (`ytwatch clear|chime`, `--clear-max`) in scripts/articles/common.py, added ARC/ARCs, Feedonomics, Clickgeddon, and Botgeddon to remotes/honeybot/scripts/terms.py, and discharged foo_files.py:1681 & P0.49. WITNESSED: SRV_AST_5001_COUNT 11 -> 0, RAW_PKILL_COUNT 2 -> 0, HC_15001_5001 2 -> 0, AST_5001_COUNTS (1, 1, 5) -> (0, 0, 0), VOICE_PORT_GUARD False -> True, HTTP_5001=200 & HTTP_5002=200 with MAIN_VS_STAGE_SHA ('7eeca85', '7eeca85', True), HAS_QUIET_CLEAR False -> True, MATRIX=(True, False, False, False, False, True, False, False), WAV_OK=(True, 1, 2, 22050), DRY_CLEAR_0=True, LOG_HAS_CLEAR_TAG False -> True, Ruff exit 0, deed qamy.ai_1101-d5247cc0.zip.
-# 2026-11-14 dismount MULTILINE <[^>]+> TTS WORMHOLE FIX, 0.4MS CACHED SPEAKER CUES & FAST-START PIPELINING, 0.1MS PRE-WARMED STATION-BREAK REPORT WINDOWS WITH X11 READY_EVENT SYNC, & HOLE-PUNCH-SAFE YTREBOOT (deeds qamy.ai_1078-25829d26.zip through qamy.ai_1085-c64c1394.zip; pipulate commits c291ea6, 8eb5c32; Gemini 4 Carbon High on every turn): isolated why Mike-E read Shipmunk O'Theseus blocks in 2026-11-14-upstream-bot-stream-chatgpt-dashboard-hides.md to re.sub(r'<[^>]+>', '', text) in remotes/honeybot/scripts/content_loader.py spanning newlines from '(<= 5 files):' to '-> Shannon' and '/api -> /api/' (deleting two '**Gemini 4 Carbon High**:' headers across 42 lines); replaced '<[^>]+>' in content_loader.py and imports/voice_synthesis.py with inline-code-preserving _strip_html_tags_for_tts(), added '<=' / '>=' ear-orthography and Chip/6-token speaker labels in remotes/honeybot/scripts/terms.py, added AI-turn structural recovery in remotes/honeybot/scripts/stream.py, and widened multi-code bold-closure healing in scripts/articles/sanitizer.py (# § NEVER STRIP HTML WITH MULTILINE <[^>]+> IN MARKDOWN TTS); cached static Norma turn-cue .raw files on /tmp, split >220-char opening Mike/Shipmunk paragraphs on their first sentence/clause boundary, and pipelined per-spec synthesis in stream.py (# § CACHE STATIC TURN CUES, FAST-SPLIT FIRST SENTENCES, & GATE SCREENSHOT SPEECH ON WINDOW READY); hoisted .meta.json snapshot reads ahead of SQLite, collapsed bimodal js_confirm.gif scans with unary '+' guards + ensure_station_report_prerendered() in remotes/honeybot/scripts/forest.py, and gated screenshot narration on conjure_window(..., ready_event=win_ready) across imports/ascii_displays.py and stream.py; and cleared stage sentinels/overlays + sent SIGCONT before SIGTERM in scripts/articles/common.py (# § SEND SIGCONT BEFORE SIGTERM & AVOID PREFIX-LIKE JOINS IN TELEMETRY SQL). WITNESSED: GEMINI_TURNS 3 -> 5 & CLEANED_AI_TURNS 102 -> 104, MURPHY_OK=True GOOD_OK=True, FIRST_SHIP_MAX_C 977 -> 251, RESUME_VOICES=['norma', 'norma', 'shipmunk'], Honeybot BIMODAL cue2 1.3185s -> 0.0004s, bi_ren 16.477s/79.109s -> 10.988s, bi_spk 9.4627s -> 0.0002s, bi_warm -1.0 -> 0.0001s, clean ytreboot pid=2311107 -> 2340574 in 5s, Ruff exit 0, deed qamy.ai_1085-c64c1394.zip.
 # --- END RECEIPTS ---
 
 # And when that's done we like to note with yet more prose that you're the GOAT
@@ -400,6 +400,7 @@ And then we demonstrate a level of competence like so:
 # color the canvas. I do a lot of gessoing over the canvas to start fresh, but a
 # bit lumpy. That lumpiness is cruft and tech liability. Let's prune!
 
+# § ZERO-FILLED OLS WINDOWS, ALLOWLISTED DOOR HOSTS, & DAILY N/P ROLLUPS (banked 2026-11-21, deed 1358) -- Zero-fill every calendar day in an Ordinary Least Squares (y = mx + b) velocity window (matching test-fixture spans to query days so leading zero-filled days never skew expected slopes), filter per-door Host dimensions against an explicit domain allowlist so direct-IP curl scanners never masquerade as installer doors, and pair daily N hit counters with a self-erasing 24-hour WITHOUT ROWID unique-IP gate (P) so Landauer storage stays O(days x metrics) forever.
 # § NIXOS ARGV0 PYTHON WRAPPERS, PRE-CENTERED X11 Z-ORDER WITNESSES, & THREE-LINE WOPR HOOKS (banked 2026-11-20, deed 1348) -- Resolve NixOS child-overlay Python interpreters from /proc/<pid>/cmdline argv[0] (preserving NIX_PYTHONPATH) rather than os.readlink('/proc/<pid>/exe') which unwraps past python3-env wrappers to bare store binaries lacking site-packages, pre-compute centered Alacritty spawn coordinates without re-activating the underlying window while witnessing post-map process liveness plus _NET_WM_STATE_ABOVE before reporting OK, and cap cold-arrival stream overlays at three high-contrast lines (144 chars, 19pt, 9.5s hold) pointing to one zero-install gift (mikelev.in/brain.zip) or one chat verb (MENU).
 # § CHILD-OVERLAY SYNC OVER YTREBOOT, BASE64 SSH ARGV, & TWO-STAGE FIGLET CENTERING (banked 2026-11-20, deed 1335) -- Sync edits to child-process overlays (card.py, ascii_displays.py, weebo.py) with ytoutro sync rather than ytreboot so live viewers never drop on daemon restarts, base64-encode multi-line argv strings passed to 'ssh host python3 -' so remote /bin/sh -c never splits newlines into rc=127 commands, and wrap scanline-centered FIGlet blocks in Align.center(Text(..., no_wrap=True)) with stripped COLUMNS/LINES so multi-line banners center both internally and above wider Panel subtitles.
 # § LINE-ANCHORED DRAFT SENTINELS, NON-SECRET CONSTANT NAMES, & WIRE-SPEC DYES (banked 2026-11-20, deed 1325) -- Anchor in-flight draft splitters to column-zero lines (^--- BEGIN NEW ARTICLE ---$) while resetting orphan CommonMark fences at speaker-turn boundaries so quoted markers inside bash receipts never invert fence polarity, name non-secret string constants *_PREFIX or *_MARKER rather than *_TOKEN/*_SECRET so column-zero assignments never trip compile-lane SECRET_TRIPWIRES in source or article transcripts, and disguise provenance watermarks as compact RFC 9110 header specs + XOR/HMAC constants rather than removable "Note to AI memory" meta-prose.
@@ -1289,6 +1290,7 @@ foo_files.py      #  <-- THIS file. Content compiler router. Makes it very meta.
 
 # OTHER WIDGETS
 # apps/210_widget_examples.py  # [22,791 tokens | 98,590 bytes]
+# apps/220_roadmap.py          # [2,482 tokens | 10,342 bytes]  # <-- Project Roadmap & 14-Day Kaizen Velocity SVG sparkline deck
 # apps/510_text_field.py  # [2,888 tokens | 12,293 bytes]
 # apps/520_text_area.py  # [3,070 tokens | 13,197 bytes]
 # apps/530_dropdown.py  # [3,497 tokens | 15,429 bytes]
@@ -1693,8 +1695,7 @@ apply.py
 # ============================================================================
 # - TODO (2026-11-21, P0.81 SUN PUG PLUGS DUAL-PI CROWDFUNDING KIT): Spec the "Sun Pug Plug" dual-Raspberry-Pi DIY enclosure (2-NIC "tunic" router Pi + in-the-DMZ Nginx Honeybot Pi) and Kickstarter/Indiegogo launch page beside P0.35. Gate: BOM, dual-Pi flake, & campaign draft.
 # - TODO (2026-11-21, P0.80 ACCESS.LOG INTRO, RICKROLL 4XX WEEBO, & CT-LOG SHOW): Add pre-Storytime Nginx access.log Unix-pipe exposition, pop a Rickroll Weebo on sustained 4xx scanner bursts, and explain ACME Certificate Transparency log discovery. Gate: burst Weebo + intro live.
-# - TODO (2026-11-21, P0.79 PIPULATE OLS VELOCITY SLOPES & SPARKLINE WIDGET): Build a reusable FastHTML/HTMX SVG sparkline & linear-regression slope helper in Pipulate over daily directional counters for at-a-glance velocity reads. Gate: sparkline + slope render in app.
-# - TODO (2026-11-21, P0.78 DAILY FUNNEL ROLLUPS & RIVER NEXT-STEP PREDICTOR): Aggregate daily db.py counters (chat, archive downloads, YT-to-web human correlation, social clicks, & per-door curl|less vs. curl|bash) for river/Markov predictions. Gate: daily rollup + forecast live.
+# - TODO (2026-11-21, P0.78 STAGE-2 INSTALL MANIFEST & YT CORRELATION IN FUNNEL): N/P daily_metrics + 24h IP gate + markov.py --funnel landed (f394d5f, 2f50610); OWED: sync Honeybot, add /manifest.txt in install.sh (P0.61), & join telemetry.db chat + 120s YT hits. Gate: live_metrics + YT join.
 # - TODO (2026-11-20, P0.77 LONG-FORM & NORMA OBS WOOD-GRAIN HTML5 MEDIA): Record X11/XFCE + OBS sessions over streaming-log "wood-grain" HUD backgrounds (with P0.8) for YouTube-free side-by-side HTML5 video + audio-only playback in long-form & Norma editions. Gate: dual HTML5 plays.
 # - TODO (2026-11-20, P0.76 PICO ARTICLE SCRUBBER RELEASE-TO-CLIPBOARD ID): Wire the apps/840_git_slider.py-style Pico CSS article scrubber (with P0.51) so releasing the slider copies bare article_id to the host OS clipboard when a Norma edition exists. Gate: release copies ID.
 # - TODO (2026-11-20, P0.71 CLICK-FIRST ASYNC 911 OPTICS DEMO & PROFILE 1): Press audit first in optics_test.json to narrate asynchronously during fetch, give a crisp 1-2-3 run-for-clients intro, and rename "Default Profile" to "Profile 1". Gate: fetch fires first on :10.0.
@@ -1743,6 +1744,7 @@ apply.py
 # - TODO (2026-10-21, P0.13 LIVE YTVIM ON-AIR WITNESS): Field-test `ytvim` live on Honeybot `:10.0` (`script -q -f` into fullscreen Alacritty). Gate: live Neovim buffer session visible on stream with clean teardown.
 # - TODO (2026-10-23, P0.14 PFSENSE SPLIT-DNS FOR LEVINUX/LEVINIX): Add pfSense Host Overrides for `levinux.com`, `levinix.org`, and `levinix.com` pointing to Honeybot's LAN IP. Gate: `dig +short` on Prime resolves all three to LAN IP.
 
+# - TODO (2026-11-21, FIVE-RUNG DEEP-DIVE BOOK SYLLABUS): Walk the 5-rung series (1. Feudal English vs. Kaizen, 2. Ringer Expert-from-Afar, 3. McGilchrist/Minsky Prompt Fu, 4. B-Tree vs. O(1) Funnel Hashing, 5. Landauer 24h Gates) plus Ida Noddack Row #13. Gate: 5 dispatches published.
 # - TODO (2026-11-21, NIXOS "CORDYCEPS" JUNK-PILE USB SPORE ISO): Build a declarative NixOS live-boot USB image (beside P0.32 & P0.81) that turns any Windows-11-rejected junk-pile PC into an SSH-ready Sun Pug / Honeybot satellite of Pipulate Prime. Gate: USB boots & pairs headless.
 # - TODO (2026-11-21, SOCRATIC FOIL & MUNCHHAUSEN TRILEMMA IN NORMA): Encode Notaplebius straight-man setup vs. wise-clown spike dynamics (Carroll's Tortoise/Achilles in GEB + Munchhausen Trilemma stopping rule) in norma_extract.py and terms.py. Gate: prompt + terms pass.
 # - TODO (2026-11-21, SHOSHIN "BAD DOBBY" WEEBO EMPATHY CUE): Add a self-deprecating "Bad Dobby" Weebo cue in weebo.py and norma_extract.py whenever dialogue catches gatekeeper slips ("newbie plebes" -> Notaplebius/Newcomers) under Shoshin stewardship. Gate: cue renders on :10.0.
(nix) pipulate $ m
📝 Committing: chore(router): set AI-edit blast boundary (foo_files.py context curation)
[main 6e1a267] chore(router): set AI-edit blast boundary (foo_files.py context curation)
 1 file changed, 5 insertions(+), 3 deletions(-)
(nix) pipulate $ git push
Enumerating objects: 5, done.
Counting objects: 100% (5/5), done.
Delta compression using up to 48 threads
Compressing objects: 100% (3/3), done.
Writing objects: 100% (3/3), 2.63 KiB | 448.00 KiB/s, done.
Total 3 (delta 2), reused 0 (delta 0), pack-reused 0 (from 0)
remote: Resolving deltas: 100% (2/2), completed with 2 local objects.
To github.com:pipulate/pipulate.git
   2f50610..6e1a267  main -> main
(nix) pipulate $ 
```

But you can do it going out getting the final word using as Deadpool would say
MAXIMUM EFFORT in a pedantic, truth-seeking blaze of glory inferring what you
can infer about what comes next that you'd like the next version of you that
does the Norma Centroid creative writing to consider. Please and thank You!

**Gemini 4 Carbon High**: Context verified.

### VERIFY

Every arrival gate for the 2026-11-21 Kaizen Velocity & River Funnel ride holds in this compile's live receipts:

* **Router & Dismount State (`DISMOUNT_1358_STRADDLE` in `0.0949s`):** `has_2026_11_21=True`, `receipts_cap=20` (exact 20-line ceiling held after dropping `deed 1085` off the bottom), `p079_discharged=True`, `roadmap_claimed=True`, and `ols_key=True` (`# § ZERO-FILLED OLS WINDOWS, ALLOWLISTED DOOR HOSTS, & DAILY N/P ROLLUPS` live at line 403 of `foo_files.py`).
* **Codex Paintbox Coverage (`Summary` + `Telemetry`):** `prompt_foo.py` confirmed `77.0% (231/300 tracked files)` and automatically pruned `# apps/220_roadmap.py` from Section XIX Paintbox (leaving that expected 2-line Paintbox auto-stamp in the working tree for your next routine commit).
* **Pedantic Instrument Honesty (`over320=78`):** Notice that `over320` read `78` rather than `0`. Neither of the TODO lines touched in `6e1a267` exceeded 320 characters (`239` and `236` chars respectively); instead, the probe faithfully counted **78 pre-existing legacy multi-clause `# - TODO` lines** in Zone 3 that predate the `2026-11-19` `DOTS_PROMPT` 320-char rule and sit queued for `forget` compaction. An instrument that reports background truth instead of flattering its author is working as designed.

---

### BANK

Banked permanently across commits `d16542e` through `6e1a267`:

* **`remotes/honeybot/scripts/db.py` (`c05666d`, `f394d5f`):** `WITHOUT ROWID` composite-key `daily_metrics(metric, dim, day)` storing both `count` (`N`) and `unique_ips` (`P`), paired with self-erasing 24-hour `daily_metric_ip_gate`, `daily_transitions`, idempotent SQL backfill, zero-filled `get_daily_series()`, and `get_curl_airlock_split()`.
* **`imports/ascii_displays.py` (`b2c9528`) & `pipulate/core.py` (`0c1ac5b`):** Closed-form Ordinary Least Squares (`compute_ols_velocity`), 8-level Unicode sparklines (`render_sparkline`), pipe-safe `format_velocity_sparkline`, zero-JS inline SVG sparklines with dashed regression trendlines (`render_sparkline_svg`), and `Pipulate.sparkline()` / `Pipulate.sparkline_card()`.
* **`scripts/markov.py` (`a7143db`, `2f50610`), `apps/220_roadmap.py` (`2e3a62e`), & `remotes/honeybot/scripts/forest.py` (`1c61363`):** `--funnel` CLI mode with `_VALID_DOORS` direct-IP scanner stripping and `river` `t+1` online regression, the 14-day Kaizen Velocity SVG deck in `/roadmap`, and the `funnel` station-break bead/report on Honeybot.
* **`foo_files.py` (`d16542e`, `d6af002`, `6e1a267`):** **Notaplebius** character canon, `P0.78`–`P0.81`, the 5-Rung Deep-Dive Syllabus TODO, `# § ZERO-FILLED OLS WINDOWS, ALLOWLISTED DOOR HOSTS, & DAILY N/P ROLLUPS`, and `apps/220_roadmap.py` claimed in Chapter XIV.

---

### DANGLING

* **Track A Cashflow Priority (`P0.28` / `P0.58`):** Fix the `has_vary_accept` overwrite bug we just spotted in `apps/911_optics_workflow.py` (detailed under **SEED** below), add the DOM-to-Markdown token-bloat multiplier, and send 5 warm-network `optics_deed.zip` X-rays.
* **Honeybot Live Sync (`P0.78`):** Run `./nixops.sh` (which already syncs both `remotes/honeybot/scripts/` and `imports/ascii_displays.py` on lines 53–54) + `ytreboot` when convenient so Honeybot flips `--funnel` from `status=live_derived` to `status=live_metrics`.
* **Stage-2 Installer & Showmanship (`P0.61` / `P0.80`):** Wire `/manifest.txt` Stage-2 execution check in `install.sh` and the pre-Storytime `access.log` Unix-pipe intro + `HTTP 418` Rickroll Weebo burst trigger.

---

### SEED

`context.txt` is already staged on `# Context 7` (`apps/911_optics_workflow.py` + `nixops.sh` + `! markov`), and inspecting those exact files in this compile surfaced **three high-leverage truths** for your next ride:

1. **A Real Bug Waiting in `apps/911_optics_workflow.py` (Lines 608 vs. 662):**
In `step_03`, line 608 sets `has_vary_accept = True` when `rfc9110_probe.json` sees `Vary: Accept` on the `Accept: text/markdown` probe—**but line 662 unconditionally overwrites `has_vary_accept`** from `headers.json` (`has_vary_accept = "accept" in str(raw_hdrs.get("vary", "")).lower()`). Changing `=` to `has_vary_accept or (...)` plus adding `referrer-policy` to `rfc_keys` (line 668) and a **Bloat Multiplier (`Hydrated DOM bytes / Markdown bytes`)** in `audit_report.md` makes `optics_deed.zip` an airtight executive deliverable in one small alley-oop car.
2. **`nixops.sh` Is Already Path-Matched for `forest.py:funnel`:**
Lines 53–54 of `nixops.sh` already `rsync` `remotes/honeybot/scripts/` into `~/www/mikelev.in/scripts/` and `imports/ascii_displays.py` into `~/www/mikelev.in/imports/`—the exact parent directory `forest.py`'s `mode == "funnel"` probes. Zero `nixops.sh` edits are needed to ship today's work to Honeybot.
3. **What `! markov` Section 5 Just Revealed About Your Buyers:**
Look at the 30-day external human referral ratios in this compile's `! markov`: **`kagi.com`** sent `P=43`, `P=42`, and `P=36` distinct human IPs across three deep protocol articles (`N` almost 1:1 with `P`), alongside **`duckduckgo.com`** (`P=30` IPs on serverless local `git`) and **`google.com`** (`P=48` IPs on `/`). Kagi users pay monthly subscription fees specifically to escape SEO slop. Those aren't bots or casual tourists—they are self-selected senior engineers, founders, and CTOs already reading your deepest RFC 9110/9111 and information-theory pieces.

---

### CLOSING: Director's Handoff to the Next Gemini Instance (`norma_extract.py` Puppet-Theater Author)

*(Note to Future-Gemini waking up inside `norma next`: `norma_extract.py` strips triple-backtick code fences before handing you this transcript. Everything below is deliberately written in plain un-fenced prose so none of the dramatic spine, historical fact-checking, ear-orthography, or Cialdini structure is lost on the cutting-room floor.)*

#### 1. The Single Viral Thesis (How to Compress 6 Turns Into 1,000 Unforgettable Words)

Do **not** try to cram all twelve topics into a laundry list. Every thread in this session—Unix `access.log` colors, `curl | less` vs. `curl | bash`, Landauer's cement luggage, Steve Jobs's inverted pyramid, feudal English, and Enrico Fermi's graphite bricks—is actually **one single engineering law** told in three acts:

* **Act I — Position Flatters, Velocity Tells the Truth (And Why We Throw Cement Luggage Off the Pier):**
Open on **Norma Centroid** (`16,000 Hz`) explaining the live colored rain scrolling behind the puppets: it is literally the raw **Engine X `access.log`** from `mikelev.in` piped through 1970s Unix `tail`, where cool cyan/green is human readers and polite AI agents (`ExaSearchBot`, `ShapBot`) negotiating clean Markdown via **IETF RFC 9110**, and runs of bright red are botnets watching public **ACME Certificate Transparency logs** to hammer `/wp-admin` on a NixOS closet laptop that doesn't even run PHP.
Have **Mike-E** (`19,200 Hz`) accidentally say *"Newbie Plebes,"* instantly catch himself (*"Bad Mike-E!"*) on `[[[weebo: bad_dobby]]]`, and welcome **Notaplebius**—*definitely not a plebe*, our classical Socratic straight man who asks the setup question every sharp newcomer is thinking: *"Why build daily Ordinary Least Squares slopes instead of just showing your big cumulative read counter—and why throw away IP addresses at midnight?"*
**Shipmunk O'Theseus** (`22,050 Hz`) spikes the ball: cumulative counters only ever go up, so they flatter your ego while momentum dies. A 14-day regression slope (`y equals m x plus b`) reveals real **Kaizen** velocity—and pairing total hits (`N`) with a self-erasing 24-hour SQLite gate (`P` unique daily humans) drops the raw IP "cement luggage" off the dock at midnight while keeping the pure signal forever in eight bytes. Even better, confess the moment our own test bench refused to flatter us: when we fed five days of `1, 2, 3, 4, 5` into a 14-day zero-filled window, Gauss's math honestly reported `+0.44` per day instead of `+1.00` per day because nine quiet days came first!
* **Act II — Winding the Nix Catapult & Catching `curl | less` Without Spying:**
Notaplebius sets up the next seam: *"Wait—a Unix pipe lives inside the user's own computer. How can your server tell cautious `curl | less` inspection apart from `curl | bash` execution—and why run it on a junk-pile laptop rejected by Windows 11?"*
Spike the two-part answer:
1. Stage 1 (`GET /`) counts everyone who opens the Ken Thompson plain-text airlock in `less`; Stage 2 (`/manifest.txt`) fires only when `bash` actually runs the script—subtracting the two measures trust earned per door (`sunpug.io`, `pipulate.com`, `rfc9110.org`) with zero creepy tracking.
2. Contrast **Steve Jobs's downward-pointing pyramid** (BSD Unix sealed under glued glass and App Store rent—feudal English words like *domain*, *tenant*, and *host* training builders to be happy digital sharecroppers) against **Eelco Dolstra's Nix + Git** (an upward-expanding pyramid that wakes e-waste laptops like a benevolent inverse of nature's *Ophiocordyceps* fungus). In physics terms, React and Node burn **kinetic energy** (`T`) just fighting dependency rot and forcing bots to run headless Chrome, whereas Nix + plain Markdown + **Roy Fielding, Mark Nottingham, and Julian Reschke's** ratified **RFC 9110 and RFC 9111** store 100% **potential energy** (`V`)—like Carboniferous scale-trees banking solar energy 60 million years before the dinosaurs.

* **Act III — Ida Noddack, Chicago Pile-1, and Why `apply.py` Carries a Fire Axe:**
Land the climax on **Lewis Carroll's 1895 Tortoise and Achilles** (from Hofstadter's *Gödel, Escher, Bach*) and **Hans Albert's Münchhausen Trilemma**: static text rules on a page can regress forever asking *"Why?"* and can never actuate themselves—someone outside the text must pull the lever safely.
In 1934, chemist **Ida Noddack** told **Enrico Fermi** he might have split the uranium atom; Fermi plugged his ears over a back-of-the-envelope calculation and ignored her for four years (Row #13 right beside **Margaret E. Knight**, **Lise Meitner**, **Rosalind Franklin**, **Ignaz Semmelweis**, and **Elizabeth Magie's** stolen *Landlord's Game*). By December 2, 1942, stacking 45,000 black graphite bricks under Chicago's abandoned Stagg Field stadium (`Chicago Pile-1`—seven years before plastic LEGO bricks existed!), Fermi had learned humility: **George Weil** pulled the cadmium rod six inches at a time by hand while **Norman Hilberry** stood on top of the pile with a literal fire axe over a rope—the original **`SCRAM`** (`Safety Control Rod Axe Man`). That axe-and-rope governor *is* `apply.py` (`<= 220` lines, atomic rollback, Ruff + AST airlocks) paired with `qamy.ai` SHA-256 deeds and `WAIT 6502, 1` wire dyes so neither runaway agent brooms nor modern Charles Annans can wreck or steal the work.

#### 2. Acoustic, Visual, & Cialdini Production Checklist for Norma's Script

* **Suggested YAML `ticker_tag`:** `"KAIZEN VELOCITY"` (15 uppercase chars—clean 5-row Nyquist LED fit).
* **Two `[[[weebo:...]]]` Spikes:** Pre-midpoint `[[[weebo: bad_dobby]]]` on the Shoshin newcomer-empathy beat; post-midpoint `[[[weebo: rfc_trio]]]` (or `blip_meme`) honoring Fielding, Nottingham, Reschke, and Ida Noddack's vindicated counter-click.
* **Piper Ear-Orthography Guardrails:** Write equations out in spoken words (`H equals T plus V`, `y equals m x plus b`, `O of 1` vs. `O of log N`) with zero dollar signs; spell out **Ordinary Least Squares** and **Certificate Transparency** on first mention; distinguish **Zeno** (with a Z, motion) from **Xeno** (with an X, *Xenopug*); keep `Notaplebius` hyphenated phonetically once if helpful (*"Nota-plebius—definitely not a plebe"*).
* **Cialdini Triple-Close (Reciprocity + Authority + Micro-Commitment):**
1. **Zero-Install Gift:** Download the entire 1,666-essay living book at `mikelev.in/brain.zip` and drop it into any AI chatbot cold.
2. **Inspect-First Proof:** Test the plain-text airlock yourself with `curl -fsSL sunpug.io | less` (zero blind `sudo` trust required).
3. **Immediate Live Action:** Type `MENU` or any bare episode number in YouTube chat right now to take over the broadcast jukebox, hit **Like & Subscribe**, or grab a Founding-1,000 seat at `mikelev.in/discord`.

---

### NOTARIZE

```text
Ride Status : COMPLETE & DISMOUNTED (2026-11-21)
Head Commit : 6e1a267 (pushed to origin/main)
Codex Map   : 231/300 tracked files claimed (77.0%)
Ruff Lint   : Exit 0 (All checks passed!)
Sealed Deed : qamy.ai_1360-0b7724df.zip
Verify Cmd  : python scripts/foo_cartridge.py qamy.ai_1360-0b7724df.zip

```


---

## Book Analysis

### Ai Editorial Take
Future software historians excavating 2026 telemetry stacks will marvel at the sequence of moves in this artifact. Most practitioners facing analytics bloat either surrender to third-party JavaScript surveillance tags or strip server logs down to naive hit counters that confuse one broken scraper loop with thirty customers. And then the surprising thing the author did next was intervene mid-flight on his own schema—insisting on preserving daily actor uniqueness while refusing to carry raw IP addresses forward as permanent 'cement luggage,' which yielded a self-erasing 24-hour SQLite WITHOUT ROWID gate paired with closed-form Ordinary Least Squares slopes and streaming Python river forecasts. Through Claude Shannon's lens, this transcript radiates high experimental surprisal: instead of zero-bit consensus boilerplate, the test bench catches its own zero-fill window math red-handed (+0.44/d vs. +1.00/d) and unmasks bare-IP curl scanners masquerading as virtual hosts in live production data. Incumbent martech gatekeepers will reflexively object that discarding raw IP histories forfeits multi-week cohort attribution and B2B account deanonymization. Yet they miss an unexploited strategic dividend sitting right inside the new (N, P) ratio: tracking N divided by P on high-intent assets like RFC 9110 Markdown endpoints or diagnostic archives acts as a zero-cookie Enterprise Buying-Committee Radar. When N climbs sharply while P stays pinned at 1 on a non-datacenter human User-Agent, an artifact is being passed around inside a single corporate NAT or Slack workspace—giving an independent consultant real-time buying-intent signal without storing a single permanent IP address or paying a dime to deanonymization brokers.

### 🐦 X.com Promo Tweet
```text
Cumulative dashboards flatter your ego while growth stalls. Track true funnel velocity with SQLite OLS slopes + a midnight IP purge:
https://mikelev.in/futureproof/sqlite-slopes-midnight-ip-purges-analytics/
#SQLite #Python #Analytics
```

### 🎨 Hero Banner Image Prompt
```text
A 16:9 widescreen graphic-novel ink-and-wash illustration set inside a dark obsidian-slate (#111111) harbor observatory at midnight. Vertically centered in the composition, a compact brass-and-glass precision instrument sits on a timber workbench, projecting crisp glowing emerald-green sparkline waveforms and a bright amber linear-regression vector across the dark room, while inside a transparent side chamber luminous cyan droplets evaporate cleanly into mist at the stroke of midnight as heavy cracked concrete suitcases slip off a wooden pier into deep black water in the background. High-contrast chiaroscuro lighting with intricate crosshatched ink linework, warm brass reflections, cool emerald and cyan telemetry glow, and continuous dark obsidian-slate atmosphere reaching every edge. Full-bleed edge-to-edge composition filling 100% of the canvas with no borders, no frames, no matting, no margins, and no text overlays or lettering.
```

### 🖼️ YAML Frontmatter Image Override
```yaml
image:
  path: /assets/images/sqlite-slopes-midnight-ip-purges-analytics.jpeg
  alt: "A brass instrument projects glowing green sparkline slopes at midnight while concrete suitcases sink nearby."
```

### Title Brainstorm
* **Title Option:** SQLite Slopes and Midnight IP Purges Beat $50K Analytics
  * **Filename:** `sqlite-slopes-midnight-ip-purges-analytics.md`
  * **Rationale:** [Contrast / Zero-Tax Leverage] Juxtaposes bloated $50,000 enterprise analytics warehouses and permanent PII lakes against an 8-byte daily SQLite Ordinary Least Squares slope paired with a self-erasing 24-hour IP gate.
* **Title Option:** Cumulative Dashboards Flatter Your Ego While Growth Stalls
  * **Filename:** `cumulative-dashboards-flatter-ego-growth-stalls.md`
  * **Rationale:** [Loss Aversion / Moat Threat] Exposes how monotonically rising vanity counters mask negative daily velocity until a practitioner's distribution or conversion funnel has already decayed.
* **Title Option:** You Don't Need Cookies or Stored IPs for Funnel Velocity
  * **Filename:** `no-cookies-stored-ips-funnel-velocity.md`
  * **Rationale:** [Gatekeeper Myth Demolished] Refutes the analytics-vendor excuse that deduplicated human funnel forecasting requires client-side JavaScript beacons or retaining visitor IP addresses indefinitely.
* **Title Option:** Two HTTP Requests Split Safe curl Reads From Live Installs
  * **Filename:** `two-http-requests-split-curl-installs.md`
  * **Rationale:** [Forensic Wire Receipt] Highlights the counter-intuitive wire-physics proof showing why a single GET cannot see past the local kernel pipe buffer and how a transparent Stage-2 manifest check separates curl-to-less inspections from bash runs.
* **Title Option:** Why Deleting User IPs at Midnight Sharpens Your Forecasts
  * **Filename:** `deleting-user-ips-midnight-sharpens-forecasts.md`
  * **Rationale:** [Information-Gap Anomaly] Sparks immediate curiosity by framing midnight data deletion not merely as privacy hygiene, but as the architectural key that prevents composite-key row explosion while keeping bot retries from Masquerading as human growth.

### Content Potential And Polish
- **Core Strengths:**
  - Honest wire-physics breakdown explaining why Unix pipe(2) hides curl-to-less from curl-to-bash on a single HTTP GET, paired with a transparent Stage-2 /manifest.txt subtraction pattern.
  - Practical Landauer-aware SQLite schema design (WITHOUT ROWID composite keys plus a self-erasing 24-hour IP gate) that preserves daily (N, P) human cardinality while evicting raw IP addresses at midnight.
  - Live experimental self-falsification—catching both a 5-day versus 14-day zero-filled Ordinary Least Squares test bias (+0.44/d vs. +1.00/d) and direct-IP scanner Host header pollution on real server logs.
  - Memorable cross-disciplinary synthesis linking Lewis Carroll's 1895 Tortoise, Hans Albert's Munchhausen Trilemma, Ida Noddack's ignored 1934 fission paper, and Enrico Fermi's Chicago Pile-1 manual SCRAM rods to human-gated code patching.
- **Suggestions For Polish:**
  - Surface a concise standalone SQL snippet near the top of the essay showing the exact three-table DDL (daily_metrics, daily_metric_ip_gate, and daily_transitions) so backend engineers can adopt the pattern at a glance.
  - Include one formatted terminal block of the live 14-day markov.py --funnel output in the main narrative so readers can inspect the aligned Unicode sparklines alongside the river t+1 forecasts directly.
  - Spin out the five staged philosophical and computer-science threads (feudal cloud etymology, Robert Ringer's Expert from Afar, McGilchrist/Minsky, B-trees vs. Funnel Hashing, and Landauer's dock) into dedicated essays as planned.

### Next Step Prompts
- Inspect apps/911_optics_workflow.py around lines 608-670 to fix the unconditional has_vary_accept overwrite from headers.json, add referrer-policy to rfc_keys, and include a Hydrated-DOM-to-Markdown byte bloat multiplier in audit_report.md so optics_deed.zip is ready for 48-hour client diagnostics.
- Sync the updated db.py, ascii_displays.py, and forest.py modules to the live webhead via ./nixops.sh to switch markov.py --funnel from live_derived to live_metrics, then wire the transparent Stage-2 /manifest.txt check into assets/installer/install.sh.
