What problem does it solve? Teams document architecture decisions and standards but rarely verify that the actual codebase follows them. This Skill closes that gap by systematically comparing the "should be" (standards, compliance criteria, AGENTS.md rules) against the "is" (real code, folder structure, dependency manifests) and producing a prioritized gap report. ## Core Features & Use Cases - Criterion-level compliance checking: Audits each verifiable compliance criterion (CR-XXX) from docs/standards/ and each atomic rule from AGENTS.md, citing the originating ADR for traceability. - Fitness function execution: Detects and runs automated architecture checks (dependency-cruiser, ArchUnit, import-linter, or a scripts/arch/verify runner) as primary evidence, and suggests new ones for automatable criteria lacking coverage. - Prioritized, revalidatable reports: Writes a report in docs/audits/arch-audit-YYYY-MM-DD.md grouped by high/medium/low priority, with support for revalidations that append changes without rewriting history. - Use Case: Before a release, ask the agent to verify the code still respects the documented standards; it greps the repo, runs the architecture checks, and delivers a report listing exact offending files with suggested actions. ## Quick Start Ask the agent to audit whether the current codebase complies with the architecture standards and AGENTS.md rules and generate the compliance report.
