repo-doctor
Run a structured whole-repository health evaluation that returns categorized findings, severity-tagged issues, category scores, an overall health score, and a prioritized remediation plan.
When this skill is meant to trigger
Run a structured whole-repository health evaluation that returns categorized findings, severity-tagged issues, category scores, an overall health score, and a prioritized remediation plan. Use this skill whenever the user wants to audit, evaluate, review, score, health-check, or 'doctor' a codebase — including phrases like 'repo doctor', 'audit this repo', 'health check this codebase', 'what's wrong with this project', 'review this codebase', 'what should I clean up', 'tech debt audit', 'code quality scan', 'is this repo production-ready', 'check for security smells', or any general request to assess the overall state of a repository. Also trigger before handing a repo to another team, when onboarding an agent loop into a codebase, or when the user asks 'what would you fix first'. Produces both a JSON report (repo-doctor-report.json) and a Markdown report. Prefer this skill over ad-hoc review when the user wants a whole-repo view rather than a single-file critique.
How this skill works
Reviews pasted code for security, quality, and cleanup priorities. Expect a ranked findings list with concrete places to look, not a rewrite of the whole repo. Use the paste field for the tree and the files that matter.
Parts of this skill expect files, a shell, or a browser. It works from your input and returns any files inline.
Fast and cheap. The default for most skills. Catalog rates $0.3 in and $2.5 out per million tokens. The receipt uses OpenRouter's billed cost.
Paste your input and run the skill. The result is written here as it streams.