Hand-written runbooks rot. They describe the system as it was the day they were written, list steps nobody has run in a year, and quietly omit the verify step because the author "knew it worked." The one document you reach for at 2 AM is the one you trust least. For an AI agent it's worse: a procedure that was imagined rather than performed has never been proven to converge, and its rollback is a guess.
So invert the source of truth. Instead of writing a runbook from memory, derive it from what actually happened. Runbook Forge reads a resource's facts and its append-only change journal — every change made, with how it was verified and rolled back — and synthesizes a runbook where every operation has been performed and verified at least once. Done five times? It says so. Verification signal recorded? It's in the steps.
The runbook earns trust because it's evidence, not aspiration.
This is the Runbooks layer of QA Veritas — a platform exploring how AI agents reason about, verify, and operate complex systems. It closes the arc from memory to executable procedure: what one agent recorded becomes what the next one safely runs.
Repo + write-up in the comments.
First comment: Repo: github.com/qa-veritas/runbook-forge · Platform: github.com/qa-veritas