A polish Charter that was supposed to be cosmetic cleanup surfaced ten production gaps in six hours — including two features that had shipped to main and never functionally worked for ten days. The anti-pattern got named, the pattern got documented as fw-4.18.0, and the CLI helper got deferred on purpose.
An agent flagged a stale spec before being asked to. The reconstruction of that morning, and the design property that surfaces it — codified after the fact as Principle #8 of StrayMark.
Twenty-seven hours later, the manual discipline got two names: Pattern 1 — pre-declare SpecKit refresh, and Pattern 2 — post-close audit-driven Batch N.4. The canonical document, the telemetry schema, the CLI helpers — and one hour after that, a meta-pattern that reabsorbs both upward.
Four months between the project's first AIDEC — "AI agents are a single technical audience" — and the open standard that operationalized it. AGENTS.md at the repo root, read by fifteen vendor CLIs, replacing the proliferation of CLAUDE.md / GEMINI.md / .cursorrules and a dozen more.
Two gestures in seventy-two hours: closing the last 5% of i18n parity (20/20/20 governance docs across EN + ES + zh-CN) and publishing the CLA badge in the README. Small separately; together, the moment the framework stops talking to itself and starts talking to whoever wants to read it.
Seven consecutive Charters on a one-month-old plan in Sentinel, executing CHARTER-18 by hand with the feeling that any false step would bury six Charters of work. Five hours to canonize the manual discipline as upstream governance — without naming the pattern yet.
One file among dozens lived backwards — Spanish-canonical in an EN-canonical framework, for six weeks, without anyone noticing. A maintenance fix that records something the blog keeps running into: outliers teach you the convention by contrast.
Zero TDEs across thirteen closed Charters in Sentinel, despite at least seven recognizable pieces of transversal debt. The document type existed; the operational trigger didn't. How fw-4.13.0 closed the gap — and how its PR chain accidentally codified a lesson about stacked PRs.
464 lines of bash got replaced with a flag. From "I check this looks like a document" to "I check this satisfies the canonical schema, in the mode you ask for, with concrete operational consequences." Phase 2 of the CLI roadmap and the formalization of validate.
Six hours of session work with a capable agent and the framework's full onboarding loaded — and not once did Charters get suggested. Five minutes to apply them when asked directly. The asymmetry isn't about capability; it's about visibility.