README Doctor
9-point README rubric with P0/P1/P2 punch list, evidence-cited and category/audience-aware.
Werkt met setup
Wat het doet
Audits a README against a 9-criterion rubric (hero presence, install-to-first-success length, the 'what is this' sentence, audience-jargon match, scannability, drift signals like stale versions and badge sprawl, contribution surface) weighted by inferred repo category and audience, then outputs a prioritized punch list with line-cited evidence. Triggers on requests to audit, review, or diagnose why a README isn't landing; explicitly does not rewrite prose or lint markdown syntax.
Testrapport
Section 1.1 ('operating mode') and the PR-opening step in Phase 4 both say 'same as repo-visuals' without defining the modes inline -- repo-visuals lives in a separate plugin (plugins/repo-visuals) not bundled with readme-doctor, so a verbatim install is missing that piece; the core 9-point audit rubric itself is fully self-contained and produced a sharp, evidence-cited, prioritized punch list on a test README.
Getest op: 2026-07-15 · Claude Code 2.x (agent harness)
Installatie
git clone https://github.com/livlign/claude-skills cd claude-skills mkdir -p ~/.claude/skills cp -r plugins/repo-doctor/skills/readme-doctor ~/.claude/skills/readme-doctor
Commando's en voorbeeldprompts
/readme-doctor9-point README rubric with P0/P1/P2 punch list, evidence-cited and category/audience-aware.
Skills reageren op gewone verzoeken — geen commando's om te onthouden. Na installatie activeren prompts zoals deze de skill (in het Engels):
why isn't my README landing with new visitors, what's missingaudit this README and give me a prioritized list of fixesdoes my project's README explain what it does fast enough