#673 · Primary category: AI Agents & Automation

exo

agent harness self-evolution

Exo is an agent + harness architecture that is fully recursive, able to safely edit all aspects of itself at runtime to get better at your tasks.

Project last updated:08/30/26

GitHub Stars

897

Forks

79

Contributors

11

License

MIT

Why we included this project

Exo is built for people who want a long-running agent that can change its own code, prompts, and tooling at runtime rather than treating those parts as fixed. The harness keeps a canonical event log of everything it has tried and uses that history to improve itself incrementally, with an eye to avoiding loops. It comes as a minimal framework with a sandboxed Docker runtime, a scheduler for recurring tasks, and chat adapters for CLI, web chat, IRC, Discord, WhatsApp, Signal, and Slack. Because it keeps durable infrastructure like state and security separate from agent-specific logic, it is a reasonable starting point when you want an agent that evolves its own behavior over time instead of staying frozen at whatever you first configured.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category