Autonomous Loops

核心功能是提供6种不同复杂度的Claude代码自动化循环模式，涵盖顺序管道、PR自动化、DAG编排等，覆盖日常任务到大型项目场景。可通过触发短语或决策矩阵选模式，输出无人工干预的迭代开发、CI/CD式流水线等流程，需依赖gh、git等工具，遵循隔离、上下文持久等原则。

paudyyin

热度 49.01

最近更新 2026-07-18 07:17:49

分类：开发依赖：代码执行 / 读写文件 / 内部调用大模型 / 第三方 API应用场景：DevOps / 自动化脚本 / 工作流 / 任务管理

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add autonomous-loops

Metadata

name: autonomous-loops
description: Autonomous Claude Code loop patterns: sequential pipelines, persistent REPL sessions, parallel spec-driven generation, PR automation, cleanup passes, and RFC-driven DAG orchestration. Choose pattern by complexity: simple (sequential) → medium (PR loop, infinite agents) → advanced (DAG with merge queue). Trigger phrases: autonomous loop, agent loop, continuous development, parallel agents, multi-pass refinement.
metadata: {"clawdbot":{"emoji":"🔄","requires":{"bins":[],"env":[]},"os":["linux","darwin","win32"]}}

Autonomous Loops — Patterns for Claude Code Automation

Running Claude in loops enables spec-driven development, CI/CD-style pipelines, and iterative refinement without human intervention between steps.

Quick Start

Choose your pattern by complexity:

Sequential Pipeline (simple) — Chain claude -p calls for linear workflows
Persistent REPL (simple) — Interactive sessions with history
Spec-Driven Parallel (medium) — Deploy N agents from spec, manage waves
PR Automation Loop (medium) — PR creation, CI fix, auto-merge
De-Sloppify Pass (add-on) — Cleanup step after any implementation
RFC-Driven DAG (advanced) — Multi-unit parallel work with dependency graph

Pattern Spectrum

Pattern	Setup	Complexity	Best For
Sequential Pipeline	Bash script	Low	Daily tasks, scripted workflows
REPL	Node/CLI	Low	Interactive development
Parallel Agents	Claude Code loop	Medium	Content generation, spec variations
PR Loop	Shell script	Medium	Iterative multi-day projects
De-Sloppify	Add-on to any	Optional	Quality cleanup after implementation
DAG Orchestration	Python/Node	High	Large features, parallel units, merge coordination

References

references/sequential-pipeline.md — Basic claude -p loops with examples
references/persistent-repl.md — NanoClaw-style session persistence
references/parallel-agents.md — Spec-driven deployment with wave management
references/pr-automation.md — Continuous Claude PR loop with CI gates
references/de-sloppify.md — Quality cleanup pattern
references/dag-orchestration.md — RFC-driven multi-unit coordination

Key Principles

Isolation — Each loop iteration gets fresh context (no bleed-through)
Context Persistence — Use files (SHARED_TASK_NOTES.md) to bridge iterations
Exit Conditions — Always set max-runs, max-cost, max-duration, or completion signal
No Blind Retries — Capture error context for next iteration
Separate Concerns — Different loop patterns for different problem sizes

Decision Matrix

Is this a single focused change?
├─ Yes → Sequential Pipeline
└─ No → Do you have a spec/RFC?
         ├─ Yes → Do you need parallel work?
         │        ├─ Yes → DAG Orchestration
         │        └─ No → PR Automation Loop
         └─ No → Do you need many variations?
                  ├─ Yes → Parallel Agents + Spec
                  └─ No → Sequential Pipeline + De-Sloppify

Anti-Patterns

❌ Infinite loops without exit conditions ❌ No context bridge between iterations ❌ Retrying the same failure without capturing error context ❌ Negative instructions instead of cleanup passes ❌ All agents in one context window (reviewer should never be the author) ❌ Ignoring file overlap in parallel work

Adapted from everything-claude-code by @affaan-m (MIT)