Most Claude Code projects run on a CLAUDE.md and not much else. The rest of the configuration surface - agents, hooks, skills, plugins, memory, permissions - goes unused because the knowledge to set it up is scattered. cc-rig resolves it: pick your framework and workflow, get 30-65 native files generated in seconds.

How it works

The tool treats configuration as two orthogonal dimensions. Framework (16 templates: FastAPI, Django, Flask, Next.js, Express, Rails, Laravel, Spring Boot, ASP.NET, Phoenix, Gin, Echo, Go stdlib, Rust CLI, Rust/Axum, generic) determines technical content — lint commands, test conventions, agent docs, MCP servers. Workflow (5 presets: speedrun, standard, spec-driven, gtd-lite, verify-heavy) determines process — how many agents, how strict the review gates, whether autonomous loops run.

Pick one of each and cc-rig resolves the full configuration: 3-13 specialized agents with scoped tool access, 6-19 slash commands, safety hooks that block destructive operations, framework-matched community skills, official plugins, and a team memory system tracked in git.

Key capabilities

  • Auto-detection: cc-rig init --migrate reads project files (pyproject.toml, package.json, go.mod, Cargo.toml, etc.) and proposes a configuration
  • Expert mode: Override defaults — pick individual agents, commands, plugins, hooks
  • Runtime harness: Optional task tracking, budget awareness with model-aware pricing, verification gates, and autonomy loops with stuck detection
  • Team sharing: Export .cc-rig.json config so teammates get identical setups
  • Safe updates: cc-rig config update re-runs the wizard with existing choices pre-filled, preserving manual edits via manifest tracking
  • Clean removal: cc-rig clean removes exactly what was generated

Philosophy

cc-rig is a one-shot generator. It writes native Claude Code files and gets out of the way. The output is the same configuration you’d write by hand if you had the time and the research. Delete cc-rig afterward and everything keeps working.