ai-project-scan
AI Project Scan
Scan an existing repo's governance and architecture files. Produce a read-only decision report: what agents currently have, what's missing or stale, and which Repo Forge skill to run next.
Purpose
Before changing anything in a repo's architecture, you need to know what you actually have.
Many repos have governance files — READMEs, plans, decision notes — but they are incomplete, outdated, duplicated, or not aligned with each other. Agents working with inconsistent context produce inconsistent work. This skill audits the current state, classifies every architecture asset, and tells you exactly what to keep, what to fix, and which Repo Forge skill to run next.
The scan is read-only. No files are changed. The output is an actionable decision report.
Primary outcome:
- inventory of architecture/governance files
- compatibility assessment vs canonical patterns
- recommendation: keep, normalize with keeper, or bootstrap with seed
- optional pattern-assimilation candidates
- intake-pattern assessment when raw request routing is present or missing
The cardinal rule: This is a read-only audit. Never modify any file during a scan — the report's value comes from seeing the repo exactly as agents will see it.