auto
Autonomous AI coding loop (Ralph Wiggum methodology). Use when running AI agents in unattended/autonomous mode to independently select and complete tasks from a structured task list (prd.json). Enables AFK development across multiple context windows.
npx skills add majiayu000/claude-skill-registry --skill auto-ar4mirez-samuel --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# Autonomous AI Coding Loop Enable unattended AI-driven development using the Ralph Wiggum methodology. **Use When**: Running AI agents autonomously, AFK development, batch task execution **4D Phase**: Develop (autonomous execution of generated tasks) --- ## Overview The autonomous loop runs AI agents in a cycle, where each iteration: 1. Starts with fresh context (no conversation history dependency) 2. Reads project state from files (prd.json, progress.md, CLAUDE.md) 3. Selects and implements one task 4. Commits changes and updates state 5. Documents learnings for future iterations Continuity is maintained through **persistent files**, not conversation history. --- ## Core Principles ### 1. Fresh Context Per Iteration Each cycle begins with a clean AI context window. The AI agent reconstructs understanding from: git history, prd.json (task state), progress.md (learnings), and CLAUDE.md/AGENTS.md (project guardrails). ### 2. Right-Sized Decomposition Tasks must be completable within a single context window. Break large features into focused, atomic steps. If a task is too big, split it before starting. ### 3. Feedback-Driven Quality Automated checks (tests, linting, type checking) s
- Overview
- Core Principles
- 1. Fresh Context Per Iteration
- 2. Right-Sized Decomposition
- 3. Feedback-Driven Quality
- 4. Knowledge Persistence
- 5. Agent-Driven Prioritization
- Workflow
- Option A: PRD-Based Setup (Planned Work)
- Option B: Pilot Mode (Zero Setup)
- Running the Loop
- Monitoring
- Manual Intervention
- Per-Iteration Protocol
Use .claude/skills/create-prd/SKILL.md Use .claude/skills/generate-tasks/SKILL.md samuel auto init --prd .claude/tasks/0001-prd-feature.md cat .claude/auto/prd.json # Task list in JSON cat .claude/auto/prompt.md # Iteration prompt Fully autonomous — no PRD or task files needed samuel auto pilot Customize iterations, focus, and discovery frequency samuel auto pilot --iterations 20 --focus testing samuel auto pilot --discover-interval 3 --max-tasks 5
What does the auto skill do?
Autonomous AI coding loop (Ralph Wiggum methodology). Use when running AI agents in unattended/autonomous mode to independently select and complete tasks from a structured task list (prd.json). Enables AFK development across multiple context windows.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill auto-ar4mirez-samuel --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.