harnessscore

Public repo scanner

How mature is your AI coding harness?

Paste a public GitHub repository URL to get a maturity level (L0–L4), dimension scores, and failing checks with remediation guidance — the same deterministic model as the CLI.

Public repos only · Max 100 MB · Results cached 7 days · 5 new scans per IP per day

Sample output

Same report as the CLI

Deterministic scoring — zero LLM calls, zero network inside the scanner. The web app downloads your repo once, then runs the same checks as npx harness-score.

Open the guide
  harness-score v1.5.3  ~/my-app

  Maturity: L2 · Guided   Score: 70/108 (65%)
  Detected: Cursor, Claude Code

  Context & Guides     ████████████████░░░░  80%
  Skills & Commands    █████████████░░░░░░░  65%
  Hooks & Guardrails   ░░░░░░░░░░░░░░░░░░░░   0%
  Sensors & Feedback   ████████████████░░░░  80%
  CI Feedback          ██████████████░░░░░░  71%
  Hygiene & Safety     ███████████████░░░░░  74%

  To reach L3: sensors ≥ 60%; ci ≥ 50%

Maturity ladder

Five levels, one score

From unharnessed agents to self-correcting loops — the same L0–L4 model as the CLI and CI gate.

L0L0 · UnharnessedCold-start every sessionL1L1 · DocumentedAGENTS.md orients the agentL2L2 · GuidedRules, skills, hygieneL3L3 · SensingTests, types, CI verify workL4L4 · Self-correctingHooks close the loop
Read the maturity model →

What you get

Diagnosis, not just a number

  • Maturity level from Unharnessed (L0) to Self-correcting (L4)
  • Six dimension scores: context, skills, hooks, sensors, CI, hygiene
  • Failing checks with evidence and docs links
  • Optional GitHub OAuth to open a PR adding a README badge