What problem does it solve? When an AI agent first encounters a repository, it lacks durable knowledge of the codebase layout, build system, and conventions, forcing repeated re-exploration on every session. This Skill performs a structured full knowledge audit of a repository and persists the findings as reusable memories and skills. ## Core Features & Use Cases - Guided repository walk: Systematically traverses top-level directories, README anchors, manifests (Cargo.toml, go.mod, package.json), build scripts, and CI workflows to map subsystems and their interactions. - Memory reconciliation: Recalls existing workspace memories first, verifies them against disk, and appends or saves typed memories (architecture, convention, api, gotcha, build) without duplicating entries. - Skill gap detection: Lists existing skills and authors new SKILL.md files only for workflows solved two or more times, then reports memories touched and the area of least confidence. - Use Case: A developer onboarding an agent to a large Rust monorepo asks for a full knowledge index; the agent maps core/, tui/, sdk/, and web/ subsystems, saves build-test-release-flow memory, and finishes with a confidence report. ## Quick Start Run a full knowledge index of this repository to map its architecture, build flow, and conventions into persistent memories.
