aeonfun/aeon

The most autonomous AI agent framework: runs unattended on GitHub Actions, self-healing skills, drives Claude Code, Grok, Codex & more. No approval loops. Configure once, forget forever.

What it solves

Aeon is an autonomous agent framework designed to perform complex tasks unattended, removing the need for constant human approval loops or "babysitting." It allows users to automate workflows like shipping code, deploying apps, and conducting security research without manual intervention.

How it works

Aeon operates as a GitHub repository combined with GitHub Actions, using a configuration file (aeon.yml) to schedule tasks via cron. It uses "skills"—which are simply Markdown files containing a prompt and frontmatter—to define capabilities. The framework supports six different agent harnesses (including Claude Code, Grok, and Kimi) and can be integrated into notification channels like Telegram or Discord.

Who it’s for

It is built for developers and security researchers who want to create self-sustaining AI agents that can run on a schedule, remember information across runs, and manage their own infrastructure.

Highlights

  • Self-Healing: Automatically detects and repairs broken skills using a health-check and repair loop.
  • Self-Improving: Can write new skills for itself based on a simple sentence.
  • Multi-Agent Orchestration: Able to spawn specialized instances (fleets) of itself to handle different domains like crypto or security.
  • Unattended Execution: Runs entirely on GitHub Actions, meaning it requires no local hosting once configured.
  • Integrated Guardrails: Includes read-only modes and an optional authorization layer (Fleet Watcher) to prevent irreversible actions.

Related

  • Project
  • Project
  • Project
  • Project
  • Project