mixxan05

Agent Loops

Name: Agent Loops
Author: mixxan05

AI community

SOURCE https://github.com/mixxan05/agent-loops

Description

Deterministic agent loops for AI coding CLIs (Claude Code, Codex, Cursor, Kimi): worker + adversarial verifier agents, hard quality gates, automatic retries, human escalation. File-based, zero dependencies, one-click on/off.

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

Agent Loop

Stop auditing every AI agent output. Define tasks and quality gates once — a deterministic loop orchestrates worker and verifier agents until the work is proven correct or escalated to you.

[![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE) [![Platform](https://img.shields.io/badge/platform-Windows%20%7C%20macOS%20%7C%20Linux-blue.svg)](#install-into-your-project) [![PowerShell 5.1+](https://img.shields.io/badge/PowerShell-5.1%2B-5391FE.svg)](#install-into-your-project) [![PRs Welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg)](CONTRIBUTING.md)

**🇩🇪 [Deutsche Version dieser Anleitung → README.de.md](README.de.md)**

Works natively with **Claude Code** and generically with **any agent CLI** (OpenAI Codex, Cursor, Kimi, custom) — same files, same protocol, swappable agent. IDE-only agents (e.g. Antigravity) are supported through a semi-automatic manual mode.

The Problem
How It Works
The Switch — On/Off by Design
Quickstart (Token-Free Demo)
Install Into Your Project
Agent Presets
Configuration Reference
File Layout
The Overview Page
Claude Code Native Integration
Writing Good Tasks and Gates
Language Note
Limitations (Honest Ones)
FAQ
Project Documentation
Contributing & License

The Problem

If you run several coding agents in parallel, **you** become the bottleneck: every output needs your review, every session needs your orchestration. Agent Loop inverts this: