What problem does it solve? Existing codebases accumulate implicit architectural decisions and living conventions that were never formally documented, making onboarding and governance difficult. This Skill analyzes a repository's structure, dependencies, and source code to surface those hidden decisions and turn them into documented ADRs and domain standards. ## Core Features & Use Cases - Architecture Discovery: Scans folder structure, dependency manifests (package.json, pyproject.toml, pom.xml, Cargo.toml, go.mod, and more), and source code patterns to infer technology choices, layering, and design patterns. - Candidate Classification: Distinguishes historical decisions (ADR-only) from living verifiable rules (requirements grouped into one of nine canonical functional domain standards such as testing, API, security, or persistence), and skips anything already documented. - Prioritized Presentation & Creation: Presents a prioritized candidate list (high/medium/low) for user approval, then delegates approved items to the arch-manage skill to create ADRs, standard requirements, and compliance criteria in the chosen architecture root. - Use Case: You join a legacy project with no architecture docs. Run the discovery to get a prioritized list like "layered architecture", "PHPUnit + Playwright testing", and "Prisma as ORM", approve the relevant ones, and have the ADRs and standards written into docs/adr/ and docs/standards/. ## Quick Start Analyze this repository and discover which architectural decisions and coding conventions deserve to be documented as ADRs and standards.
