What problem does it solve? Before an AI agent can plan, debug, fix, or review code, it needs fast and accurate context about the codebase. Manually exploring project structure, hunting for relevant files, and tracing dependencies wastes time and context tokens. This Skill performs read-only codebase reconnaissance and returns a structured report so downstream skills can act with full context. ## Core Features & Use Cases - Structure and Framework Detection: Maps directory layout and detects the stack from config files like package.json, Cargo.toml, pyproject.toml, go.mod, and pom.xml. - Targeted Search with Saturation Detection: Runs focused glob and grep searches for a caller's domain, capped at 10 file reads, and stops early when new reads stop yielding new entities. - Dependency and Convention Mapping: Traces import relationships to identify blast radius, and detects linting, naming, and test framework conventions. - Search-First Assessment: Classifies each request as ADOPT, EXTEND, COMPOSE, or BUILD so callers know whether an existing solution already exists. - Use Case: A debugging skill is stuck on a failing auth module. Scout produces a zoom-out Mermaid diagram showing the target file, its sibling modules, and its callers, so the agent can see the full picture before proposing a fix. ## Quick Start Ask the agent to scan the codebase and report the project structure, framework, and files relevant to the authentication module.
