Journal Archive — Page 15 of 165
2026-10-02
The Door Table and the Moving Twig: Replayable Installs and the Stick Bug Secret
Decouple installer domains with dynamic door tables, embed stealth telemetry via Operation Stick Bug, and turn AI claims into checkable offline receipts.
2026-10-01
The Scrollback Rule: Engineering Full-Width Terminal Dividers for Replayable Logs
A single-line bash trick creates full-width scrollback dividers before clear -x, simplifying log extraction and replayable receipts in the Age of AI.
2026-10-01
The Prefix Ladder and the Unalias Guard: Replayable Shell Ergonomics in the Age of AI
Refactoring shell hooks with command prefix ladders and unalias guards to create frictionless, checkable single-letter workflows for rapid human-AI turns.
2026-09-30
Forever-Commands and the Talk Toggle: Replayable CLI Habits in the Age of AI
Learn how short Unix-style commands build lasting muscle memory, and see why verifiable proof straddles catch last-inch shell execution gaps.
2026-09-29
Layered Reality and Operator Overloading: Reading the Physics of CPython in the Age of AI
How David Deutsch's layered reality, CPython's ceval.c, and operator overloading shape an anti-agentic system that keeps humans learning in the Age of AI.
2026-09-29
The Three-Door Installer: Named Prompts, Split DNS, and Verifiable Releases
Fixing pfSense split DNS for a third domain led to named terminal prompts, audit-ready installer pipes, and synchronized multi-door releases.
2026-09-29
The Deed Is Named for the Door: Engineering Replayable QA Archives for AI
Replace opaque prompt cartridges with self-explaining, replayable QA zip archives. Learn how verifiable receipts turn AI vibes into reliable audit trails.
2026-09-29
The Qamy Door and the Cellular Witness: Launching a Third Domain with Replayable Receipts
Deploying qamy.ai from a home NixOS server using the proof straddle, exact-match AI patches, and verifiable cellular access logs without opaque containers.
2026-09-28
The Forward-Slash Test: Bridging Agent Skills to Claude Code with Verifiable Receipts
How a single tracked symlink turns open Agent Skills into native Claude Code commands while keeping tool calls anchored in verifiable local receipts.
2026-09-28
AGENTS.md and Agent Skills: Refactoring Pipulate for a Checkable PyPI Release
A practical refactor aligns Pipulate with AGENTS.md and Agent Skills, moves its journal index on demand, and ships agent-facing files in PyPI.