#238 · Primary category: AI Agents & Automation

evolver

a2a agent-evolution agent-framework agent-protocol ai-agent auditable-ai autonomous-agent cli evomap gep llm-agent mcp memory-system nodejs prompt-engineering prompt-governance self-evolution self-evolving skill-library skills

The GEP-powered self-evolving engine for AI agents. Auditable evolution with Genes, Capsules, and Events. | evomap.ai

Project last updated:08/25/26

GitHub Stars

9.0K

Forks

840

Contributors

9

License

GPL-3.0

Why we included this project

Most agent frameworks grow by having you hand-edit prompts when things go wrong. Evolver takes a different route: it treats an agent's accumulated experience as compact, versioned 'Genes' and 'Capsules' managed under the GEP protocol, and logs every change as an auditable event. You run it as a Node.js CLI inside a git repository, point it at your memory directory, and it picks a gene and produces a prompt your host runtime can consume; single-run, review, and continuous-loop modes cover a one-off improvement, inspecting changes before accepting them, or letting evolution run on its own. Because those genes live in git, each improvement is traceable and reversible, so agents can evolve without quietly drifting past what you actually want. It runs fully offline, with an optional EvoMap network connection for sharing skills across agents, which makes it as useful for a solo side project as for a team standardizing how its agents grow.

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