stale-claims
Skill2 starsChanged 2 days agoMITView on GitHub
Monthly upkeep of dated claims, external links and model ids. Use for stale doc, dated claim, model id, monthly improve.
In this repository4 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- plan-slicesSkill
- verify-changeSkill
The file
Files
House skill — owned by this repository's maintainers, and written from this repository's own evidence. Correct it in place; setup never overwrites a skill that exists.
- Act only on the first Monday of the month:
[ "$(date -u +%u)" = 1 ] && [ "$(date -u +%-d)" -le 7 ]. fail -> stop, file nothing, open nothing. - Dated claims:
grep -rn 'As of 20' README.md docs-> a claim older than three months that a source now contradicts is worth acting on. - External links:
git ls-files '*.md' | xargs grep -ohE 'https://[^) >]+' | sort -u, thencurl -sIL --max-time 10each -> a 404 or a domain gone is a finding; a transient failure is not. - Model ids and versions: the same grep for
claude--> a model id or tool version no longer current is a finding. - Pick at most three, in this order: broken link or contradicted claim -> stale doc to delete (with its index rows) -> compression drift from the writing rules in
CLAUDE.md. - Ship one small PR per improvement, its evidence in the body,
bash bin/checkgreen first. fail -> stop: do nothing rather than open an empty PR.
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.

