CLI for agents
Harness Score
A CLI, run with npx harness-score, that rates a repository's AI coding harness from L0 to L4 on a 108-point checklist and lists what to fix next.
Why it is useful
For teams that keep rules, skills, hooks and guardrails for Claude Code or Cursor in a repo and want a deterministic score to track or gate in CI. The README says it makes no LLM calls and no network requests.
Facts
- Kind
- CLI for agents
- Source
- https://github.com/paladini/harness-score
- Repository
- paladini/harness-score
- Author
- paladini
- License
- MIT
- Works with
- claude-code, cursor
- Added
- 2026-11-06
- Last verified
- 2026-11-06 via git ls-remote (HEAD 293bc5648e)
- Found via
- github-search