continuous-agent-loop

Orchestrate continuous autonomous agent loops with quality gates and recovery controls.

Updated Mar 24, 2026
One-click install
npx skills add https://github.com/Oruga420/claude-code-skills --skill continuous-agent-loop-oruga420
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: continuous-agent-loop
Source: https://github.com/Oruga420/claude-code-skills/tree/main/continuous-agent-loop
Command: npx skills add https://github.com/Oruga420/claude-code-skills --skill continuous-agent-loop-oruga420

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides patterns and controls to run continuous autonomous agent loops that avoid churn, enforce quality gates, and recover from repeated failures so teams can maintain measurable progress without unchecked cost or merge queue stalls.

Core Features & Use Cases

  • Loop Selection & Orchestration: Choose between RFC decomposition, strict CI/PR control, exploratory parallel generation, or sequential execution to match project needs.
  • Quality Gates & Eval Integration: Integrates quality gates and eval harnesses to prevent regressions and automate acceptance criteria before merges.
  • Recovery & Auditability: Freeze, audit, reduce scope, and replay failing units with explicit acceptance criteria to repair loops and prevent repeated retries.
  • Use Case: Run a production autonomous loop that decomposes large RFCs into tasks, continuously evaluates outputs against test gates, and automatically recovers failing components to keep the merge queue flowing.

Quick Start

Use the continuous-agent-loop skill to run a persistent multi-agent loop that decomposes a feature RFC, applies quality gates, runs eval cycles, and recovers failed units.

Frequently Asked Questions about continuous-agent-loop

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

FAQPage Schema
How do I prevent autonomous agent loop churn and merge queue stalls?

To prevent autonomous agent loop churn and merge queue stalls, orchestrate continuous agent loops with quality gates, eval integration, and recovery mechanisms that freeze, audit, and replay failing units.

What is the best way to decompose large RFCs into automated CI tasks?

The best way to decompose large RFCs into automated CI tasks is using a continuous agent loop that breaks down feature RFCs into automated CI tasks, applies quality gates, and runs eval cycles before merges.

Can I use continuous agent loops for exploratory parallel generation?

Yes, you can use continuous agent loops for exploratory parallel generation by selecting the appropriate loop orchestration mode to match your project's generation and evaluation requirements.

How do I recover from repeated failures in autonomous agent loops?

To recover from repeated failures in autonomous agent loops, apply explicit acceptance criteria, reduce scope, audit the failing units, and replay them to repair the loop and prevent unchecked retries.

Do I need eval harnesses to enforce quality gates in CI/PR workflows?

Yes, you need eval harnesses to enforce quality gates in CI/PR workflows, as they integrate with continuous agent loops to prevent regressions and automate acceptance criteria before merging code.