codebase-readiness

Codebase Readiness Assessment

Run an Agent-Ready Codebase Assessment — a scored evaluation of how well a codebase supports autonomous AI agent work, framed against the Stripe benchmark of 1,000+ AI-generated pull requests per week.

Work through the following phases sequentially.

---

Phase 1: Codebase Reconnaissance

Execute the reconnaissance script to gather project metadata before launching assessment agents. Run directly — no agent needed.

bash scripts/recon.sh

The script is located at scripts/recon.sh relative to this skill's directory.

After reviewing the output, format a Codebase Snapshot:

## Codebase Snapshot: [Project Name]

- **Primary language/framework**: [detected]
- **Language tier**: [statically-typed | dynamically-typed | gradually-typed]
  - Statically-typed: TypeScript, Go, Java, Scala, Rust, C#, Kotlin
  - Dynamically-typed: Ruby, Python (unannotated), JavaScript/Node.js, PHP
  - Gradually-typed: Python with mypy/Pydantic, TypeScript with strict:false
- **Commit count**: [X]
- **Contributors**: [X]
- **Source files**: [X]
- **Test files**: [X] (ratio: X%)
- **CI/CD**: [platform(s) found or none]
- **CLAUDE.md**: [present at path, X lines / absent]
- **Linting config**: [tools found or none]
- **Quality gates**: [copy the `Suggested Gate Maturity Level: Lx -- ...` line from recon output, e.g. "L0 -- no complexity, duplication, or dead-code tooling detected"]
- **README**: [present, X lines / absent]

The Quality gates line is required. recon.sh ends its === QUALITY GATES === section with Suggested Gate Maturity Level: Lx -- <reason>. Copy that level and reason onto the snapshot's Quality gates line. Then read references/quality-gates.md and check the suggestion against the detailed evidence above it (L0 none, L1 report-only, L2 threshold, L3 regression-aware blocking, L4 governed); change the line only when the evidence contradicts the heuristic, and say why. Do not launch the assessment agents with a snapshot that lacks this line -- the agents score only their own slice of the gate evidence and read the level from the snapshot; if it is missing, each agent re-derives it independently and the four dimensions drift apart.