context-hygiene
Skill12 starsChanged 18 days agoMITView on GitHub
Audit and repair the context layer. Use when asked to clean up context, before a quarterly review, or when the knowledge graph feels stale. Finds missing frontmatter, broken references, orphan files, and stale content, then fixes them file by file.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- AlomoesDataScience/claude-skillstheir own version, first seen Nov 16, 2026
- emanueledenaro/claude-skillstheir own version, first seen Oct 15, 2026
- wrg32786/aigent-ostheir own version, first seen Oct 13, 2026
- m2ai-mcp-servers/claude-skillstheir own version, first seen Oct 11, 2026
- Leostruka/devin-bundletheir own version, first seen Aug 15, 2026
- lightarktech/founder-coding-skillstheir own version, first seen Jul 26, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- wrg32786/aigent-os★ 26same name
- AlomoesDataScience/claude-skills★ 1same name
- lightarktech/founder-coding-skills★ 1same name
- emanueledenaro/claude-skills★ 0same name
- Leostruka/devin-bundle★ 0same name
- m2ai-mcp-servers/claude-skills★ 0same 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 repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/manifest-guardrails.mdcCursor rule
- llms.txtllms.txt
- cadenceSkill
- capture-feedbackSkill
- new-executionSkill
- new-motionSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 0 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Mechanical lint: Run npm run lint:context. Fix every error (missing title/description, broken references: entries, broken wikilinks or …
- Very specific: line 26
Contradictions: Read global/ and icp/ together and flag claims that conflict (two different ICPs, outdated competitor claims in …
- Very specific: line 37
Never invent facts to fill a template.
“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Context hygiene pass
The context layer is only useful if agents can trust it. Run this audit periodically or when onboarding a new team member.
Steps
- Mechanical lint. Run
npm run lint:context. Fix every error (missingtitle/description, brokenreferences:entries, broken wikilinks or relative links). Errors block CI; do these first. - Orphans. The linter warns about files nothing references. For each orphan, decide: link it from the file that should own the relationship, or delete it if it no longer earns its place. Do not leave orphans standing.
- Staleness. The linter warns about files untouched for more than 90
days. For each, check it against reality: recent won/lost deals in
context/client/, current pricing, live positioning. Update the file or add a dated note explaining why it is still accurate. - Contradictions. Read
global/andicp/together and flag claims that conflict (two different ICPs, outdated competitor claims inalternative/). Surface contradictions to the team in the PR description rather than silently picking a side. - Ship it. One PR titled
context: hygiene pass YYYY-MM-DD, with a summary table: files fixed, orphans resolved, stale files refreshed, contradictions flagged.
Rules
- Never invent facts to fill a template. An honest gap beats a plausible lie;
leave the section empty with a
TODO(owner):marker. - Preserve frontmatter fields you do not understand.
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.

