← ClaudeAtlas

wiki-healthlisted

Run the combined wiki health dashboard — use for a comprehensive quality overview; runs wiki-lint + wiki-staleness + wiki-gaps in sequence and outputs wiki/health.md; do NOT run those sub-skills manually in the same session or you will double-execute the lint suite.
draakaap/lorewerk · ★ 1 · AI & Automation · score 74
Install: claude install-skill draakaap/lorewerk
# Wiki Health — Combined dashboard > **Dependency note:** If any of `wiki-lint`, `wiki-staleness`, or `wiki-gaps` > already ran this session, read the existing reports (`lint-report.md`, > `staleness-report.md`, `gaps.md`) rather than re-running. The 12-check lint > suite is the most expensive sub-step — back-to-back re-runs are the most > common source of wasted work here. > **Scheduled execution (recommended):** run this skill weekly, unattended, via > your platform's scheduler (Windows Task Scheduler / cron) invoking `claude -p` > with a prompt like "run the wiki-health skill and apply its remediations". > Lesson from the field: without a schedule, reports go 20+ days stale and > remediations get recommended repeatedly but never executed. Let the weekly run > also EXECUTE the remediations the reports recommend (staleness bumps, index.md > sync, size-guard rotation). Manual invocations stay valid alongside it. ## On activation Run three checks in sequence and combine results: ### 1. Run wiki-lint Execute the full 10-check lint suite. Read the results from `wiki/lint-report.md`. ### 2. Run wiki-staleness Check for overdue pages. Read results from `wiki/staleness-report.md`. ### 3. Run wiki-gaps Review knowledge gaps. Read from `wiki/gaps.md`. ### 4. Read review queue Read `wiki/review-queue.md` for provenance status. ### 5. Generate health dashboard If the combined score is above 90%, spawn one sonnet Challenger auditor (adversarial prompt: strict binary pass/fail,