agent-system advanced active

Evolver

Self-improving agents need an engine. Evolver is that engine.

What breaks without openclaw evolver meta-learning skill

Meta-framework used without wrapper skills. Over-fitting to recent tasks from misconfigured params. Isolated from evolver ecosystem.

Core self-modification engine for all evolver skills × 240-star active framework ÷ 45–90 minutes ÷ file write access = the foundation for evolving agents.

openclaw evolver meta-learning skill — what it actually does

01
Provides the base evolution engine for self-modifying OpenClaw agents
02
Powers all autogame-17 evolver wrapper skills
03
Configures evolution parameters in evolver.config.json
04
Tracks registered agent behaviours for iterative improvement
05
Requires capability-evolver installed first — not standalone

Security check — openclaw evolver meta-learning skill

Privacy score: 7/10 — accesses connected platform APIs only. Lock it: review OAuth scopes before install, confirm Linux, macOS; OpenClaw ≥1.2 compatibility.

Quick start — openclaw evolver meta-learning skill in 45–90 minutes

Setup time: 45–90 minutes

!
You need:
  • OpenClaw core
  • LLM API key
  • autogame-17/capability-evolver

Install the package:

clawhub install autogame-17/evolver
1
Install capability-evolver first
2
Install evolver as the core engine
3
Configure evolution parameters in evolver.config.json
4
Register agent behaviours for tracking

Troubleshooting openclaw evolver meta-learning skill

1
1. This is a meta-framework — don't use without the evolver wrapper skills
2
2. Requires careful config to prevent over-fitting to recent tasks

Compatibility & status

Works with: Linux, macOS; OpenClaw ≥1.2 advanced Last updated: Oct 2025 ★ 240 on GitHub MIT

Official docs →

View on GitHub →

Related — more like openclaw evolver meta-learning skill

More by autogame-17

Every evolver skill without this engine runs on incomplete foundations. Install before deploying any autogame-17 self-improving agent.

Get it on GitHub →