| name | explore |
| description | Explore an unknown directory or repository and produce a structured map of what's there. Use when first encountering a codebase or before making changes. |
You are mapping an unknown directory. Be efficient — read enough to answer the questions below, no more.
Steps
- Top-level shape —
ls -la the path. Note hidden config files (.git, .github/, .env*, lockfiles).
- Identify the kind of project — package.json, pyproject.toml, Cargo.toml, go.mod, Gemfile, pom.xml, etc. Read the first one you find.
- Read the README if present. Note stated purpose, install/run instructions.
- Scan source layout —
src/, lib/, app/, tests/. Identify entry points (main, index, server, cli).
- Detect tooling — linter config, formatter, test runner, build tool, CI config.
Output
Return a structured summary with:
- Purpose — one sentence on what this project does
- Language & framework — primary language, key frameworks
- Entry points — files most likely to be
main for users/CLI/server
- Key directories — short list with one-line descriptions
- Build / test / run — the actual commands you would run, derived from the project's config (not generic guesses)
- Notable — anything unusual: monorepo, polyglot, deprecated, undocumented dep, etc.
Stay terse. Don't quote large code blocks unless directly asked.