Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
By marcusgoll
Autonomous deep development loop with multi-agent BUILD phase. Fresh context per task, parallel execution, intelligent retries. Includes deep-add, deep-execute, deep-plan, deep-triage, deep-fullstack skills.
npx claudepluginhub marcusgoll/deep-loop-pluginComprehensive backend testing including API, security, and database. Called by review agent.
Agent from deep-loop plugin
Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality. Focuses on recently modified code unless instructed otherwise.
Agent from deep-loop plugin
Agent from deep-loop plugin
Cancel active deep loop session. Use when user asks to 'stop deep', 'cancel deep', 'abort loop'. Cleans up state files.
Stop ralph mode PRD processing. Use when user asks to 'cancel ralph', 'stop autonomous mode'. Halts background execution.
Add tasks to deep loop queue (.deep/tasks.md). Use when user asks to 'add task', 'queue this', 'deep add'. Interactively gathers task details.
Clean up stale deep loop files. Use when user asks to 'clean up', 'remove .deep directories'. Removes old session artifacts.
Process queued tasks from .deep/tasks.md. Use when user asks to 'execute queue', 'process tasks', 'run queued work'. Supports multi-session claim-based coordination with git conflict handling.
Uses power tools
Uses Bash, Write, or Edit tools
Share bugs, ideas, or general feedback.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
loophaus — Control plane for coding agents
Automated development loop and ad-hoc workflow generation for orchestrating AI coding agents
Evidence-Driven Development Protocol — 24 command-equivalent skill surfaces plus a skill-native deep-work entry alias (cross-platform: Claude Code skills + Codex / Copilot CLI / Gemini CLI / SDK), 3-layer architecture with Skill-based phase dispatch, computational enforcement (worktree guard + phase transition injection), TDD enforcement, and receipt validation
Complete collection of battle-tested Claude Code configs from an Anthropic hackathon winner - agents, skills, hooks, and rules evolved over 10+ months of intensive daily use
Production-grade engineering skills for AI coding agents — covering the full software development lifecycle from spec to ship.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Context packing and anti-duplication guardrails
v9.0.0 | Autonomous development loop for Claude Code
A deterministic, self-correcting development protocol with multi-agent BUILD phase, TDD enforcement, and senior dev mode.
# Install from GitHub
claude plugins install marcusgoll/deep-loop-plugin
# Or add to your Claude Code settings manually
cd ~/.claude/plugins/marketplaces/local/plugins/
git clone https://github.com/marcusgoll/deep-loop-plugin.git deep-loop
/plugins refresh# Start a deep loop session
/deep implement user authentication
# Quick mode for small fixes
/deep quick fix typo in README
# Check status
/deep-status
# Cancel active session
/cancel-deep
PLAN → BUILD → REVIEW → FIX → SHIP
# Enable Task Sync for crash recovery
export DEEP_LOOP_TASKS_ENABLED=true
{
"mode": "internal", // "internal" (default) or "external"
"buildMode": "multi-agent", // "multi-agent" (default) or "single"
"maxParallel": 3, // Max concurrent task agents
"maxIterations": 10 // Safety limit
}
This codebase will outlive you.
Every shortcut becomes someone else's burden. Every hack compounds into technical debt. Fight entropy. Leave the codebase better than you found it.
| Skill | Description |
|---|---|
/deep | Full deep loop with planning |
/deep-quick | Quick mode (3 iterations, no planning) |
/deep-status | Rich status display |
/cancel-deep | Cancel active session |
/start-ralph | PRD-driven autonomous mode |
/cancel-ralph | Cancel ralph mode |
MIT