cc-settings
Claude Code configuration for the Darkroom team — installs agents, skills, hooks, and coding standards into ~/.claude/.
---
Features
- One-command install — drops shared agents, skills, hooks, and standards into ~/.claude
- Subagents & skills — 10 specialized subagents and a curated, auto-invocable skill library
- Composed settings — settings.json assembled from modular config: permissions and hooks. The MCP fragment is installed to
~/.claude.json, the only file Claude Code loads user-scope MCP servers from - Non-destructive — existing permissions, custom hooks, and local overrides survive re-installs
- One-command rollback — restore the previous backup if anything looks off
- Tamper detection — a fingerprint plus audit guard the hooks against supply-chain attacks
- Open standard — AGENTS.md is read by Codex, Cursor, Copilot, and Windsurf too
---
Install
macOS / Linux:
bash <(curl -fsSL https://raw.githubusercontent.com/darkroomengineering/cc-settings/main/setup.sh)Windows (PowerShell):
powershell -ExecutionPolicy Bypass -c "irm https://raw.githubusercontent.com/darkroomengineering/cc-settings/main/setup.ps1 | iex"Both one-liners clone the repo and run the installer. To pass flags (--light, --dry-run, …), clone first and run the bootstrap from the checkout: bash setup.sh --light or .\setup.ps1 --light.
Requires Bun ≥ 1.2.21 and git — the bootstrap installs Bun automatically if missing. The full profile also installs jq if missing; no runtime hook shells out to it (all hooks are TypeScript) — it's for the jq one-liners used in team-knowledge remediation runbooks. Re-installs are non-destructive: existing permissions, custom hooks, and local overrides survive.
Restart Claude Code after install.
Light profile (for newcomers)
New to Claude Code and don't want the full surface? Install the light profile — raw Claude Code with only two additions: the statusline and the share-learning skill.
bash setup.sh --light # macOS / Linux
.\setup.ps1 --light # Windows (PowerShell)No custom CLAUDE.md, agents, rules, profiles, MCP servers, hooks (beyond the statusline), or effort overrides — just vanilla Claude Code so you're not overwhelmed. Re-run bash setup.sh without --light any time to upgrade to the full config; both tiers are permanently supported. See MANUAL.md for the full comparison.
---
What gets installed
~/.claude/
├── AGENTS.md # Portable coding standards (read by all AI tools)
├── CLAUDE.md # Claude-Code-specific config
├── settings.json # Composed from config/*.json (permissions, hooks, MCP)
├── agents/ # 10 specialized subagents
├── skills/ # 39 auto-invocable skills
├── profiles/ # Stack contexts: nextjs, react-native, tauri, webgl, maestro, react-router
├── rules/ # Path-conditioned rules (load on-demand by file type)
└── src/ # Hook + script implementations (TypeScript)Repo → install mapping:
| Repo dir | Installs to | |----------|-------------| | agents/ | ~/.claude/agents/ | | skills/ | ~/.claude/skills/ | | rules/ | ~/.claude/rules/ | | profiles/ | ~/.claude/profiles/ | | config/*.json | ~/.claude/settings.json (composed) |
---
Common commands
bash setup.sh # Install / update (full profile)
bash setup.sh --light # Light profile: raw CC + statusline + share-learning only
bun src/setup.ts --rollback # Restore the previous backup
bun src/setup.ts --dry-run # Preview what would change
bun run compose # Preview the composed settings.json
bun run lint:skills # Validate skill frontmatter + cap (≤40)
bun run audit:hooks # Classify hooks as trusted/unknown/suspicious
bun run typecheck # TypeScript check
bun test # Run all tests---
Docs
| Doc | What's in it | |-----|-------------| | MANUAL.md | Every skill — how to invoke it, what it does | | AGENTS.md | Coding standards and guardrails (source of truth) | | CLAUDE.md | Claude-Code config, delegation rules, effort levels | | docs/profiles.md | Stack-specific profiles (Next.js, RN, Tauri, WebGL, React Router) | | docs/skill-authoring.md | Writing new skills | | CHANGELOG.md | Release history |
---
Why
Shared config means every engineer on the team runs the same agents, the same guardrails, and the same coding standards — no per-machine drift. AGENTS.md is the open standard for AI coding instructions and is also read by Codex, Cursor, Copilot, and Windsurf, so the rules follow you across tools.
---
darkroom.engineering · MIT
