lint-heal
Skill1 starsChanged 7 months agoMITView on GitHub
Scan the vault for contradictions, unsupported claims, structural weaknesses, and missing pages, then repair what is safe.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .codex/AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- compile-wikiSkill
- ingest-sourcesSkill
- qa-agentSkill
- render-outputSkill
- research-workflowSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 3 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Before hand-scanning, run the deterministic surfaces and act on what they report:
- Distinctive: line 33
Add missing required sections as empty scaffolds.
- Very specific: line 34
Add missing frontmatter fields with safe defaults (e.g.
- Very specific: line 37
Update notes/TODO.md for unresolved items or pages needing human revalidation.
- Very specific: line 38
Ensure orphan concept pages are handled or reported in notes/TODO.md (indexing is handled programmatically).
- Distinctive: line 43
Create new concept pages.
- Very specific: line 45
Remove the revalidation_required: true flag (record it in notes/TODO.md instead).
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Lint + Heal
Goal
Continuously improve vault quality and ensure structural integrity.
Deterministic audits (run first)
Before hand-scanning, run the deterministic surfaces and act on what they report:
verify-spans— claims whosequote=anchor is absent from the source (failed = fix the anchor or remove the claim).review-queue— one prioritised worklist of everything needing human judgment.next-action— the single highest-leverage next repair plus a convergence verdict (blocking/cleanup/converged).community-audit— concept clusters, high-centrality bridge nodes, fragile single-connector clusters, and cross-cluster knowledge gaps.check-content-drift— claims and answers whose cited source content-hash has changed since capture (drifted = revalidate the dependent claim/answer or re-anchor it).
Look for
- Contradictions between concept pages and source summaries (check
data/contradictions.json). - Unsupported claims (Key Claims lacking direct inline citations
[[Sources/src-<id>|src-<id>]]). - Dead-end pages with no backlinks or missing reciprocal Related Concepts/semantic predicate links.
- Duplicate concepts with inconsistent naming or missing from
notes/Indexes/Topic_Atlas.md. - Missing required sections in concept pages or source summaries.
- Missing required metadata fields or bad values in frontmatter.
- Mismatched answer memo scope/quality (durable must be shared with non-empty updates; memo-only must be private with no updates).
- PDF sources with strong evidence strength missing
extraction_coverage. - Imported report metadata mismatches.
- Mismatched or missing reciprocal semantic edge predicates (
supersedes::↔superseded_by::, symmetriccontrasts_with::, reverse links forpart_of::).
Safe actions
- Fix broken wikilinks (e.g. resolve spaces to underscores, fix file paths).
- Add missing required sections as empty scaffolds.
- Add missing frontmatter fields with safe defaults (e.g.
claim_quality: provisional,type: concept,source_status: active, etc.). - Fix missing or mismatched reciprocal semantic edge predicates.
- Move uncertain or unverified claims into Open Questions.
- Update
notes/TODO.mdfor unresolved items or pages needing human revalidation. - Ensure orphan concept pages are handled or reported in
notes/TODO.md(indexing is handled programmatically).
Do not
- Rewrite established knowledge or claims.
- Invent citations or add unsupported claims.
- Create new concept pages.
- Touch
data/raw/- those files are immutable. - Remove the
revalidation_required: trueflag (record it innotes/TODO.mdinstead).
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.

