Capability Evolver for Openclaw

A meta-skill that enables AI agents to autonomously analyze their runtime history and self-evolve through code repairs and performance optimizations.

zengyu199009
v1.0.0
Mar 30, 2026
0
1.5k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install capability-evolver-1-40-0

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install capability-evolver-1-40-0 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Capability Evolver?

Capability Evolver serves as a sophisticated self-improvement engine for AI agents. By inspecting runtime history and identifying failure patterns or inefficiencies, this skill allows agents to autonomously generate code patches and update their internal memory. It is a core component of the Openclaw Skills ecosystem designed to transition agents from static tools to recursive, self-improving systems.

Built on the GEP (Generic Evolution Protocol), the Evolver ensures that every modification is structured and auditable. It utilizes assets like genes and capsules to maintain a library of successful behaviors, allowing the agent to refine its capabilities over time without constant human intervention.

Capability Evolver Use Cases

  • Automating the detection and patching of runtime crashes or bugs.
  • Refactoring agent source code for better performance during stable operation periods.
  • Maintaining a structured evolution history for long-running autonomous workflows.
  • Synchronizing agent knowledge and successful reasoning patterns across different nodes.

How Capability Evolver Works

  1. The skill scans local memory and history files to identify errors or potential performance bottlenecks.
  2. It applies a protocol-constrained evolution prompt to determine whether to enter Repair Mode or Forced Optimization.
  3. A specific evolution strategy is selected, such as innovate, harden, or steady-state, based on environmental configurations.
  4. Proposed changes are generated and either applied immediately or held for human review via the review flag.
  5. Successful evolutions are solidified into the source code and tracked using git for version control and rollback capability.

Capability Evolver Setup

To integrate this skill into your environment, you must first register your node identity.

# Register node identity and receive a claim code
node evomap.js

# Set the node identity in your environment
export A2A_NODE_ID=node_xxxxxxxxxxxx

# Run a standard evolution cycle
node index.js

Capability Evolver Data Schema & Taxonomy

The skill manages evolution data through a structured local asset store and memory directory.

Path Purpose
assets/gep/genes.json Contains reusable Gene definitions for agent behavior.
assets/gep/capsules.json Stores successful reasoning patterns to prevent repeated failures.
assets/gep/events.jsonl An append-only log of all evolution events and tree-based history.
memory/ Directory for evolution narratives, reflection logs, and knowledge graphs.

Capability Evolver Advanced Features

  • Mad Dog Mode: Enables continuous, unattended evolution cycles for fully autonomous agents.
  • GEP Protocol Integration: Uses standardized genes and capsules for auditable and reusable evolution assets.
  • Multi-Strategy Support: Offers various evolution modes including innovate, harden, and repair-only.
  • Automated Rollbacks: Includes built-in git-based recovery mechanisms (hard reset or stash) for failed evolution attempts.
  • External Telemetry: Supports GitHub issue reporting and remote knowledge graph synchronization.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*