FREE AUDIT — 5 SPOTS

Let me scan your repo.

Running Cursor, Claude Code, or Codex on a real codebase? I'll run sourcebook against your repo, map where agents are most likely to miss files or break conventions, and send you a written brief. No cost. Just honest feedback in return.

WHAT_THE_AUDIT_COVERS

●

Co-change risk map

Files that always change together but have no import relationship — the invisible dependencies agents miss every time.

●

Hub file blast radius

Which files are imported by the most other files — the ones where agent edits have the widest downstream impact.

●

Anti-pattern signals

Reverted commits, rapid re-edit patterns, circular dependencies — what agents keep re-trying that your team already ruled out.

●

Convention gaps

Naming patterns, import conventions, test structure — what the codebase expects that isn't written down anywhere.

●

Fragile files

Files with the highest edit frequency — hard to get right, most likely to produce agent-generated regressions.

●

Written brief

A short document — what we found, what it means for your AI workflow, what to watch for. Yours to keep and share internally.

WHAT_WE_ASK_IN_RETURN

Honest feedback on what the audit found — whether it was useful, what it missed, what it got wrong. A 20-minute call if you're open to it. That's it.

We're running 5 audits, not 500. If your repo isn't a good fit we'll say so. If you don't hear back within 48h, spots are full.

REQUEST_AN_AUDIT

Public or private repo. TypeScript, Python, or Go. Teams using Cursor, Claude Code, Codex, or similar.