autonomous-loops

Automate Claude Code loops across sequential pipelines and RFC-driven DAGs.

3|2|Updated Mar 8, 2026
One-click install
npx skills add https://github.com/agentmatters/mullai-bot --skill autonomous-loops-agentmatters
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: autonomous-loops
Source: https://github.com/agentmatters/mullai-bot/tree/main/src/Mullai.Skills/Skills/claude-code-everything/autonomous-loops
Command: npx skills add https://github.com/agentmatters/mullai-bot --skill autonomous-loops-agentmatters

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Patterns and architectures for autonomous Claude Code loops — from simple sequential pipelines to RFC-driven multi-agent DAG systems.

Core Features & Use Cases

  • Provides a consolidated set of loop patterns (Sequential Pipeline, NanoClaw REPL, Infinite Agentic Loop, Continuous Claude PR Loop, De-Sloppify Pattern, Ralphinho / RFC-Driven DAG Orchestration).
  • Offers architecture guidance, per-pattern tradeoffs, and example scaffolds for implementing autonomous Claude Code workflows with clear separation of concerns and per-stage context windows.
  • Supports end-to-end lifecycle: research, planning, implementation, testing, review, and merge/land stages, with robust context bridging and eviction recovery strategies.

Quick Start

Provide an initial autonomous loop implementation using the Sequential Pipeline pattern to demonstrate a simple, repeatable workflow.

Frequently Asked Questions about autonomous-loops

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I orchestrate multi-agent Claude Code workflows with isolated context windows?

Multi-agent Claude Code orchestration uses distinct per-stage prompts and isolated context windows across separate worktrees to automate planning, implementation, testing, and review. This ensures safe, auditable progress with provenance tracking across distributed agents.

What are the best autonomous loop patterns for Claude Code pipelines?

Autonomous loop patterns for Claude Code pipelines range from simple Sequential Pipelines to RFC-driven DAGs. Other patterns include NanoClaw REPL, Infinite Agentic Loop, Continuous Claude PR Loop, De-Sloppify, and Ralphinho orchestration, each offering distinct architecture tradeoffs.

How do I set up an RFC-driven DAG system for distributed agent automation?

RFC-driven DAG systems for distributed agent automation require per-stage prompts, isolated context windows, separate worktrees, and an eviction- and merge-aware workflow with provenance tracking. The Skill provides architecture guidance and example scaffolds for implementation.

Can I automate planning, implementation, testing, and review across distributed Claude Code agents?

Automating planning, implementation, testing, and review across distributed Claude Code agents is supported end-to-end. The workflow maintains distinct context per stage with robust context bridging and eviction recovery strategies to ensure safe, auditable progress.

How do I recover context when an autonomous Claude Code loop gets evicted or interrupted?

Eviction recovery in autonomous Claude Code loops relies on eviction- and merge-aware workflows with provenance tracking. Context bridging strategies restore state across stages, ensuring interrupted loops resume safely without losing auditable progress.

When should I use a Sequential Pipeline versus a multi-agent DAG for Claude Code automation?

Sequential Pipelines suit simple, repeatable autonomous workflows, while multi-agent DAGs fit complex, RFC-driven development requiring distributed agents. The Skill provides per-pattern tradeoffs to help choose the appropriate loop architecture for your scale.