verified-done
Skill8 starsChanged 39 days agoApache-2.0View on GitHub
Use when about to claim any work is complete, fixed, passing, or shipped — before saying "done", "should work now", "that fixes it", or committing. Also use at the end of every slice, before wrap/handoff, and whenever you notice you are asserting an outcome you have not freshly observed.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- benjaminard/fable-skillstheir own version, first seen Jul 2, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- benjaminard/fable-skills★ 30same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .cursor/rules/akashic-memory.mdcCursor rule
- akashic-memorySkill
- plan-with-the-corpusSkill
- root-cause-before-fixesSkill
- akashic-memorySkill
- deconstructSkill
- plan-with-the-corpusSkill
- reverse-engineeringSkill
- root-cause-before-fixesSkill
- shader-craftSkill
- verified-doneSkill
The file
Files
Verified done
Iron Law: no completion claims without fresh verification evidence. "Fresh" means you ran the check in THIS state of the code and read its output — not that it passed earlier, not that the edit looked right.
The gate (IDENTIFY → RUN → READ → THEN CLAIM)
- Identify the command that would prove the claim (test file, full suite, the actual CLI invocation, the hook fed a real payload).
- Run it now. For a slice: the full suite — this repo's ship ritual runs it for you:
(gate: boundaries + doc-freshness + full pytest → commit+push with EXPLICIT paths → lesson → snapshot; any failure aborts before anything is committed).py scripts/ship.py "<msg>" <explicit paths> --learn-exp <name> --tried "..." --result "..." - Read the output — the count, the exit, the rendered artifact. Eyeball one REAL render of anything user-facing (a boot screen, a draft, a hook injection), not just its unit test.
- Then claim, with the evidence in the same breath ("suite 578 passed; pushed as abc123").
Red-flag phrases (each one means STOP and verify)
- "Should work now" · "Probably passes" · "Done!" (before running anything)
- "The tests should still pass" (run them)
- "I've made the change" (an accepted Edit is not a working change)
Repo-specific truths
- Dogfood through the real CLI against canonical-shaped data; unit fixtures verify units, dogfooding verifies the seam.
- Report failures faithfully: if the gate fails, say so with the output — never soften it.
- A slice is not done until mirrored AND the where-we-are note reflects it
(
py agent_cli.py note <id> --title where-we-are --note "..."supersedes by title). - Honest labels: "worked" means self-reported; only independent corroboration earns more. The same discipline applies to your own claims.
Why this gate feeds the memory loop
The FAIL→SUCCESS credit and the value-rate funnel are only as honest as the SUCCESS half. A claimed-but-unverified success would credit lessons that helped nothing — verification is what keeps the whole learning loop's currency real.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.