What problem does it solve? Governance artifacts like ADRs, RFDs, and design docs decay silently — missing alternatives, unstated assumptions, pending sign-offs — and nobody notices until a bad decision resurfaces. This Skill audits existing decision records in bulk and reports exactly where each one falls short. ## Core Features & Use Cases - Six-axis quality rubric: Scores each artifact on alternatives, assumptions, reversibility, blast radius, options depth, and status/provenance hygiene. - Parallel fan-out review: Dispatches one read-only reviewer agent per artifact in waves of up to six, collecting PASS/FAIL verdicts into a single aggregate report. - Ranked report output: Writes a per-artifact verdict table plus detailed Critical findings to .mz/reviews/, without ever editing the artifacts themselves. - Use Case: Run it over docs/decisions/ before a quarterly architecture review to find which ADRs lack human sign-off or never recorded rejected alternatives. ## Quick Start Ask the AI to review the ADRs in docs/decisions and report which ones fail the governance quality audit.
