| name | move-doctor |
| description | Use when finishing a Sui Move feature, fixing a bug, before committing Move code, or when the user types /movedoctor, asks to scan, triage, or clean up Move diagnostics. Covers Move Book conventions, Move 2024 idioms, ability mistakes, testing style, and security best practices. |
| version | 0.1.0 |
Move Doctor
Scans Sui Move codebases for convention, idiom, ability, testing, and security issues. Outputs a 0–100 health score.
After making Move code changes
Run npx move-doctor@latest --verbose --diff and check the score did not regress. If it dropped, fix the regressions before committing.
For a full cleanup pass
Run npx move-doctor@latest --verbose (without --diff) to scan the whole codebase. Fix issues by severity — errors first, then warnings, then info.
/movedoctor — full triage workflow
When the user types /movedoctor, says "run move doctor", or asks for a full triage / cleanup pass, fetch the canonical playbook and follow every step (fetching it means updates ship without a skill reinstall):
curl --fail --silent --show-error \
--header 'Cache-Control: no-cache' \
https://move.doctor/prompts/move-doctor-agent.md
It's a scan → triage → fix → re-score loop that edits the working tree directly (never commits, never opens PRs). Every finding in --json carries its own fixHint and citation — fix straight from those; there's nothing else to fetch.
If the fetch fails (offline / site down), fall back to: run move-doctor --verbose, fix errors first (security/* and abilities/* findings are real vulnerabilities, not style), apply each finding's fixHint, and re-run until the score stops rising. Never silence a finding unless you can explain why the surrounding code is a documented exception.
Command
npx move-doctor@latest [path] --verbose --diff
