What problem does it solve? AI-generated code can grow faster than a team's ability to understand, review, and maintain it. This Skill establishes technology-neutral governance rules, quality gates, and review protocols so projects gain AI's implementation speed while humans retain ownership, verification, and accountability for changes. ## Core Features & Use Cases - Four Operating Modes: Draft minimal governance rules, run read-only project audits with evidence-graded findings, implement quality gates progressively (observe before blocking), and review AI-produced plans or changesets with blocking/non-blocking/unverified findings. - Eight-Layer Quality Framework: Evaluates direction, boundaries, ownership, change shape, verification, integration, delivery readiness, and long-term maintenance, with a five-state control maturity model from "declared" to "verified effective". - Risk-Tiered Requirements: Scales information and approval demands from low-risk cosmetic fixes to critical changes involving identity, funds, or irreversible data transforms. - Use Case: A team adopting AI coding assistants asks for an audit; the Skill inspects CI configs, ownership mappings, and verification entry points, then reports which controls actually block risky changes versus which are merely documented, with gaps ranked by risk. ## Quick Start Ask the AI to perform a read-only audit of the current project's AI development controls and list existing protections, gaps ranked by severity, and minimal improvement recommendations.
