Lines vs main
225
+184 / −41 · 6 files · 11 hunks
Hatterdash
for what's top of mind
Measuring the felt: reading the diff, blame, coverage and dependencies from GitHub…
Sample dashboard — no GitHub token required
acme/ledger-api · m.vale · main ← feat/webhook-receipts
Retries were unbounded after the billing cutover. This change caps attempts, stores receipts, and adds a coverage gate on the worker.
225
+184 / −41 · 6 files · 11 hunks
62%
114 of 184 added lines were last written by commits that declare an AI author
75.7%
103 of 136 executable added lines ran in CI
75/100
strong · a score from the factors below, not a probability
| Starting point | 50 | |
| CI | all 4 passing | +25 |
| Coverage of added lines | 75.7% (ci-artifacts) | +5 |
| Tests changed | 2 of 3 changed source files | +5 |
| Security findings | none high or medium | 0 |
| Known vulnerabilities | 1 high | -10 |
| Score | strong | 75 |
Counts lines last written by commits that declare an AI author or co-author. AI use that commits don't declare isn't counted, and neither are lockfiles or generated files.
Measured from coverage.
Tests changed for 2 of 3 changed source files (by file name, not coverage).
tough-cookie Prototype Pollution vulnerability
Fixed in 4.1.3. Was 4.1.3 before this change.
fetch("http://metrics.acme-corp.io/v1")
| File type | Files | Added lines | Checked for |
|---|---|---|---|
| JavaScript | 5 | 172 | code execution, SQL injection, unescaped HTML, weak hashes on secrets, broken ciphers, cleartext HTTP, known token formats, possible credentials, world-writable or wildcard permissions |
Not scanned: 1 file (lockfile, vendored or generated, 12 lines).
Sample mode cannot mutate GitHub. Authorize to approve, comment, or close a live PR.