From david-skills
Cleans up stale documentation and plan debris left by agentic coding work. Deletes obsolete docs, converts future work into issue trackers, and preserves human-facing guides.
How this skill is triggered — by the user, by Claude, or both
Slash command
/david-skills:stale-docs-cleanupThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Delete stale docs. Git history is the archive. Code is the source of truth. Docs are for humans.
Delete stale docs. Git history is the archive. Code is the source of truth. Docs are for humans.
Use short, sharp review prose. Make decisions. Spend context on evidence and edits.
bash path/to/skill/scripts/audit_stale_docs.sh . > .scratch/stale-docs-audit.md
Section A — Issue tracker.delete, move to issue, rewrite, or keep.Pick where future work lives for this repo.
Default posture:
gh issue create.glab issue create.gh.glab..scratch/<feature>/.For local markdown issues, use this shape:
# <issue title>
## Problem
<current user-facing or developer-facing problem>
## Scope
<smallest useful change>
## Acceptance criteria
- <observable outcome>
- <test or validation signal>
Delete these when found:
docs/archive, docs/old, docs/deprecated, and similar dead-doc foldersPreserve history through git and an existing CHANGELOG.md. Create or update a changelog only for human-readable release history.
Keep docs that help humans:
Contracts describe interfaces. Strip implementation journeys, phases, owners, and future work.
For useful docs with stale language, rewrite to current state:
Use references/report-template.md for final reports. Load references/doc-hygiene-rules.md when a cleanup decision is ambiguous.
npx claudepluginhub thedavidweng/skillsSpawns a doc-maintainer agent to audit all project documentation for correctness, completeness, freshness, and consistency. It autonomously fixes issues like typos, stale code examples, and broken links, surfacing anything requiring user approval.
Audits codebase documentation for accuracy, completeness, and freshness by comparing against code structure. Auto-fixes small discrepancies in fix mode, reports structural changes. Works with any language/framework.
Scans projects for maintenance issues starting with stale docs, picks one, fixes it in a git worktree, self-reviews with competing agents, and opens a PR. Use to tidy codebases.