ryanportfolio/githelp
52.1 Early
- Commit
85cbf48279c2edbe83fece55fe4384b653c2f7ea- Collected
- 2026-07-18T18:47:02.806Z
- Sample
- 75 commits · 27 merged PRs
- Publication
- Requires prior owner approval: rule
Dimensions
| Dimension | Score | Basis |
|---|---|---|
| Agent attribution & provenance Can every commit be traced to who or what made it (a linked account, a bot identity, or a co-author trailer)? | 75/75 commits have clear provenance (linked account, bot identity, or co-author trailer). Agent-involved share is reported as context, not judged. | |
| Recorded review coverage Of the merged pull requests, how many carry a recorded review by anyone other than the author? | 0/27 merged PRs received at least one review not authored by the PR author. | |
| Recorded review catch rate When a recorded review happened, did it change anything (new commits after the review, before the merge)? | Could not verify: no reviewed PRs in sample. | |
| Human merge gate Are merges performed by people, or does automation merge by itself? | 27/27 sampled merges were performed by a non-bot account. | |
| CI test/eval gate Do automated tests exist, and did pull requests actually pass them before merging? | No CI workflow files detected. | |
| Batch size Are changes small, reviewable chunks, or thousand-line dumps? | Median merged-PR diff size is 52 lines (additions+deletions). Small batches score higher. | |
| Lead time (first commit → merge) How long does work take from first commit to being merged? | Median lead time from first PR commit to merge is 0.1h. | |
| Plan-before-code evidence Do pull requests link to an issue or plan that existed before the code? | 9/27 merged PRs reference a linked issue or plan document in the PR body. | |
| Audit-trail completeness Could a stranger reconstruct why each change happened (real descriptions, recorded reviews, recorded checks)? | Mean of: substantive PR description ≥120 chars (27/27), review recorded (0/27), CI checks recorded (2/27). |
Could not verify
Owner attestation
Stated by the repo owner. The tool has not verified this and it earns no score credit.
Independent review for this repo happens as fresh-context AI audit sessions outside GitHub (handoff audits). Those sessions leave no GitHub artifacts, so they earn no credit in the recorded-review dimensions above.
What this score means
Process discipline, not code quality
This score measures process discipline, not code quality. The tool never reads the code. It reads only GitHub metadata: commits, pull requests, reviews, check runs, and merge events. It scores only what is recorded on GitHub; discipline that leaves no artifact there earns nothing.
One question
It answers one question: when AI agents help write the code, what does this repo's history prove about the checks standing between a change and the main branch?
Scoring and grades
Each dimension is a 0 to 100 answer to one concrete question, or 'could not verify' when the evidence is missing. The overall score is the weighted average of the verified dimensions, and grades band it: 90+ Elite, 75+ Strong, 60+ Developing, 40+ Early, under 40 Ad-hoc.