Skill Evolve is an automated experience solidification engine that converts successful task executions and troubleshooting steps into reusable skill modules.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skill-evolve
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install skill-evolve 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 Skill Evolve?
Skill Evolve functions as a self-evolution manager for AI agents, designed to bridge the gap between one-off task completion and long-term knowledge retention. By systematically capturing workflows from non-trivial tasks, it ensures that once a complex problem is solved or a specific environmental hurdle is cleared, the logic is codified into a permanent SKILL.md file for immediate future reuse.
This skill is fundamental for developers looking to build a high-performing autonomous ecosystem. It prevents the AI from falling into the same 'pitfalls' twice and eliminates the need for expensive re-reasoning from zero. By integrating Skill Evolve, your agent becomes smarter with every interaction, building a localized library of Openclaw Skills tailored to your specific infrastructure and preferences.
Skill Evolve Use Cases
- Solidifying complex workflows that require 5 or more tool calls to complete.
- Documenting specific environment configurations, such as non-standard SSH ports or unique database encoding requirements.
- Recording 'lessons learned' after a user corrects an initial AI mistake to prevent future recurrence.
- Capturing multi-step deployment sequences into a single, repeatable protocol.
- Archiving troubleshooting paths once a difficult bug has been successfully resolved.
How Skill Evolve Works
- Monitoring: The system monitors agent activity for 'evolution triggers,' such as successful completion of complex tasks or finding a solution after multiple failed attempts.
- Decision Logic: A built-in decision tree evaluates if the task is repetitive, non-trivial (2+ steps), and sufficiently difficult to warrant a formal skill.
- Extraction: The agent extracts the core procedure, successful commands, and identified pitfalls from the session history.
- Formalization: A new SKILL.md is generated using the
writetool or an existing one is updated using theedittool within the structured skills directory. - Deployment: The newly created Openclaw Skills are stored in categorized directories for the agent to reference in subsequent sessions.
Skill Evolve Setup
To begin using Skill Evolve, ensure your workspace directory is initialized. You can manually check or manage your library of Openclaw Skills using the following commands:
# List all currently evolved skills
ls ~/.openclaw/workspace/skills/
# View a specific solidified skill
cat ~/.openclaw/workspace/skills/<category>/<skill-name>/SKILL.md
# Remove an obsolete or outdated skill
rm -rf ~/.openclaw/workspace/skills/<category>/<skill-name>/
Skill Evolve Data Schema & Taxonomy
Skill Evolve organizes data into a clean, hierarchical taxonomy to ensure rapid retrieval and high readability.
| File Component | Description |
|---|---|
| Frontmatter | Contains the YAML metadata including name, description, version, and triggers. |
| Procedure | A step-by-step numbered list of the successful execution path. |
| Pitfalls | A list of known failure modes and how the agent corrected them. |
| Verification | Specific criteria or commands to verify the task was completed successfully. |
Directory Structure:
devops/: Deployment and environment maintenance.debug/: Error correction and troubleshooting logic.workflow/: Multi-tool chain sequences.learned/: General experience and user-corrected preferences.
Skill Evolve Advanced Features
- Incremental Patching: Uses granular editing to update specific sections of a skill (like adding a single pitfall) without overwriting the entire procedure.
- Self-Pruning Logic: Recommends deletion or refactoring when a skill becomes obsolete or exceeds 200 lines to maintain high execution speed.
- Dynamic Triggering: Automatically recognizes when to invoke the evolution process based on tool-call density and successful resolution of user corrections.
- Custom Naming Conventions: Supports standardized naming patterns like
<tool>_<brief>or<env>_<issue>for enterprise-grade organization of Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
HP Product Support and Insights
A specialized AI skill for retrieving comprehensive HP product specifications, technical support, and retail information.
ASUS Information Guide
An intelligence skill providing AI agents with comprehensive data regarding ASUS corporate history, core business sectors, and global market positioning.
Skillweave SDK Publish Smoke Test
A specialized skill designed to verify and validate the end-to-end publishing workflow for Skillweave and ClawHub client integrations.
Business Meal Assistant (fb-business-meal-skill)
A comprehensive AI assistant for managing corporate dining workflows, ensuring policy compliance, and automating reimbursement processes.
Nintendo Intelligence & Industry Insights
A specialized knowledge skill providing deep historical context, hardware milestones, and financial data for Nintendo as a global gaming leader.
Vibe-Switch
Vibe-Switch acts as a tmux-like orchestrator for AI coding agents, allowing you to run multiple LLM-powered tools simultaneously on isolated Git worktrees.
Featured*
