Evolver
ActiveDescription
The GEP-Powered Self-Evolution Engine for AI Agents — enables agents to autonomously optimize and evolve using Genome Evolution Protocol for continuous capability improvement.
Key Features
- GEP-powered self-evolution engine encoding agent experience as Genes and Capsules for auditable evolution
- CLI tool that scans memory/logs, selects genes, and outputs GEP prompts for agent improvement
- Integration with Cursor, Claude Code, and custom MCP clients via setup-hooks command
- EvoMap network for skill sharing, worker pools, and evolution leaderboards across agents
- Continuous loop mode running as background daemon for autonomous iterative improvement
- Research-backed approach: Gene representation outperforms documentation-oriented skill packages in benchmarks
Use Cases
💡 Iterative improvement of AI agent performance through gene-based evolution
💡 Shared skill and knowledge exchange across agent networks via EvoMap
💡 Automated optimization of agent reasoning through structured experience encoding
💡 Continuous agent evolution with audit trails for compliance and debugging
Categories
Quick Start
Install: npm install -g @evomap/evolver. In any git repo, run: evolver. For review mode: evolver --review. For continuous loop: evolver --loop.