← ClaudeAtlas

repo-discoverylisted

Use when the user says 'explore this repo', 'what does this codebase do', 'give me an overview of this repo', 'onboard me to this project', or opens work in an unfamiliar repo before making changes. Read-only: surface scan, entry points (README/CLAUDE.md/ARCHITECTURE.md), source tree, key files.
oliver-chase/OliverCode · ★ 0 · Code & Development · score 68
Install: claude install-skill oliver-chase/OliverCode
# Repo Discovery — Read-Only Repository Exploration ## Process Sequential commands, path-scoping, raw output-first. Never edit during discovery. ### Step 1 — Surface Scan ```bash ls <repo-root> ``` Identify top-level structure: config files (package.json, Cargo.toml, setup.py), directory layout, README presence. ### Step 2 — Entry Points Look for: - `README.md` — project description - `CLAUDE.md` — agent instructions (the only filename Claude Code auto-loads) - `ARCHITECTURE.md` — architecture doc - `CONTRIBUTING.md` — contribution rules - `package.json`, `Cargo.toml`, `go.mod`, `Gemfile` — language + dependencies - `.env.example` or `.env.local.example` — env vars needed - `Makefile`, `Justfile`, `Taskfile.yml` — build/run commands ### Step 3 — Source Tree ```bash find <repo-root>/src -maxdepth 3 -type d | sort # or app/, lib/, core/ ``` Identify module boundaries. Read the key entry point file. ### Step 4 — Key File Scan Read these if they exist: - `scripts/` directory listing - `docs/` directory listing - Main entry point (index.js, main.py, main.rs, app.tsx, etc.) - Test directory listing ### Step 5 — Summary Produce a compact table: | Aspect | Finding | |--------|---------| | Language | `<runtime + version>` | | Entry | `<main entry point path>` | | Tests | `<test dir>` — `<runner>` | | Data | `<store + access method>` | | Config | `<env example path>` | | Docs | `<the docs that actually describe current state>` | | Agent rules | `CLAUDE.md` present / ab