What problem does it solve? Arriving at an unfamiliar repository often leads to reading files in search-result order and mistaking that for understanding. This Skill provides a structured orientation workflow that establishes the codebase's shape, real module boundaries, entry points, and one traced execution path before any deep reading begins. ## Core Features & Use Cases - Structured six-step orientation: measure size and languages, find real module boundaries via imports, read entry points, trace one path end to end, locate its tests, and record findings. - Tool-agnostic with acceleration paths: every step has a baseline using git, grep, and manifests, plus an accelerated variant for code-graph servers (e.g. architecture-overview or flow queries). - Red-flag guardrails: explicit stop conditions prevent describing architecture from directory names, skipping the traced path, or treating generated code as authored. - Use Case: A developer joins a new project and asks "walk me through the architecture" — the Skill produces a written orientation naming entry points, one traced request path, its tests, and the regions left unread. ## Quick Start Ask the agent to help you understand this repository and walk you through its architecture using the explore-codebase workflow.
