Skill Extractor for Openclaw

A specialized tool that identifies successful multi-step workflows from AI conversations and preserves them as standardized, searchable skill documents.

livangy
v1.0.0
Apr 15, 2026
0
717
0

Install & Download

1. ClawHub CLI

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

npx clawhub@latest install niuniu-skill-extractor

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 niuniu-skill-extractor 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 Extractor?

The Skill Extractor is a powerful utility designed to create a continuous learning loop for AI agents. Inspired by the Hermes Agent Skill Documents mechanism, it analyzes dialogue history to identify successful procedures, rare solutions, and complex multi-step workflows. By formalizing these experiences into a structured format, it ensures that your AI doesn't just solve a problem once, but learns the optimal path for all future occurrences.

Built for the Openclaw Skills ecosystem, this tool acts as an automated technical writer for your agent. It monitors task completion and prompts the user to save high-value sequences, effectively building a local library of expert knowledge that grows more robust with every interaction.

Skill Extractor Use Cases

  • Capturing complex multi-step workflows that involve several tool calls and logical branches.
  • Preserving solutions to rare technical bugs or unique edge cases discovered during a session.
  • Standardizing repetitive procedures into high-success-rate templates for future use.
  • Building a personalized knowledge base of Openclaw Skills tailored to specific developer environments.

How Skill Extractor Works

  1. Task Analysis: The system monitors the conversation and triggers a detection event when a task exceeds 5 rounds, uses multiple tools, or results in a clear success.
  2. Extraction: The LLM reviews the interaction history to isolate key steps and critical decision points.
  3. Standardization: A standardized SKILL.md file is generated following the agentskills.io specification.
  4. User Verification: The generated document is presented to the user for confirmation and potential refinement.
  5. Persistence & Indexing: The skill is saved to the local workspace and indexed via FTS5 full-text search for immediate retrieval in subsequent tasks.

Skill Extractor Setup

To enable the Skill Extractor within your environment, ensure your workspace directory is initialized. The skill will manage its own repository under the following path:

# Standard storage path for extracted documents
mkdir -p ~/.openclaw/workspace/skills/skill-extractor/skills/

No additional manual configuration is required as the extractor integrates directly with the Openclaw Skills runtime to monitor session state.

Skill Extractor Data Schema & Taxonomy

The Skill Extractor organizes data using a flat-file structure with metadata headers for efficient indexing. Each skill is stored as a Markdown file with the following schema:

Field Description
Frontmatter Contains the unique skill name and a brief 'When to use' description.
Procedure A numbered list of executable steps derived from the successful task.
Trigger Conditions Specific context or keywords that prompt the agent to suggest this skill.
Notes Supplementary information regarding error handling or environmental variables.

Skill Extractor Advanced Features

  • FTS5 Integration: Utilizes SQLite's FTS5 for high-performance full-text search across your entire skill library.
  • Autonomous Suggestions: Automatically prompts the user to save a skill after detecting a high-complexity, high-success interaction.
  • Contextual Matching: When starting a new task, the agent automatically searches for and suggests relevant Openclaw Skills based on the current goal description.
  • Self-Improvement Loop: Enables the agent to evolve its own capabilities based on actual user interactions and verified outcomes.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*