What problem does it solve? Teams often lack an accurate picture of how their codebase is actually structured: the documented architecture drifts from what the imports implement, dependency cycles and skip-layer violations accumulate unnoticed, and security boundaries go unmapped. This Skill reads the real structure and produces a verifiable architecture report. ## Core Features & Use Cases - Component and boundary mapping: Builds a directed component graph from workspace manifests, assigns each component a responsibility statement, and reads the contract artifacts (protocol crates, seam traits, event types) at load-bearing boundaries. - Deviation detection: Finds dependency cycles, skip-layer imports, grab-bag modules, duplicated concepts, and unjustified hubs, each stated relative to the intent the codebase itself implies. - Security boundary mapping: Records where untrusted data enters, sandbox and process-isolation mechanisms, network confinement, escape hatches, and open defaults as structural findings. - Use Case: Point it at a repository with a Sokrates _sokrates analysis to produce an onboarding-ready architecture report with validated findings rendered into an interactive HTML explorer. ## Quick Start Analyze this repository's architecture using the Sokrates analysis in the _sokrates folder and write the findings report.
