From ralph-starter
Initializes Ralph Playbook scaffolding in projects for autonomous coding loops, creating AGENTS.md, prompts, specs dir, and plan tracker. Auto-detects Node.js, Python, Rust, Go for validation setup.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ralph-starter:ralph-initThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Initialize Ralph Playbook scaffolding in a project directory.
Initialize Ralph Playbook scaffolding in a project directory.
Use ralph_init with the project path to create:
AGENTS.md — agent configurationPROMPT_plan.md and PROMPT_build.md — workflow promptsspecs/ directory for specificationsIMPLEMENTATION_PLAN.md — task trackingAuto-detects project type (Node.js, Python, Rust, Go) and configures validation commands.
After init, add your spec files to specs/, then use /ralph-plan to generate an implementation plan.
npx claudepluginhub multivmlabs/ralph-starter --plugin ralph-starterInitializes Ralph Loop infrastructure from approved Claude plans by creating .ralph/ with prd.json, loop.py, CLAUDE.md for autonomous multi-step execution.
Sets up Ralph Wiggum autonomous coding loop with bash iteration, Claude prompts, test backpressure, git commits, and GitHub backup in any directory.
Documents Ralph plugin arguments (--quick, --commit), execution modes (normal/quick), state management (.ralph-state.json), commit behavior, task loops, and error handling.