A specialized automation tool to distill architectural patterns, design tokens, and technical methodologies from any codebase into reusable Openclaw Skills.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install extraction
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 extraction 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 Pattern Extraction?
The Pattern Extraction skill is an advanced analysis engine designed to transform legacy or existing codebases into structured, actionable intelligence. It systematically scans repositories to identify intentional design decisions, unique architectural flows, and customized configurations, filtering out generic defaults to capture the true essence of a project's DNA.
By leveraging this tool within the ecosystem of Openclaw Skills, developers can bridge the gap between working code and reusable knowledge. It does not just copy files; it synthesizes design systems—including colors, typography, and spacing—UI patterns, and deployment workflows into standardized documentation and modular agent skills that can be applied to future projects seamlessly.
Pattern Extraction Use Cases
- Identifying and documenting a legacy project's design tokens and aesthetic direction for future UI consistency.
- Converting a complex internal architecture into a reusable methodology for onboarding new team members via Openclaw Skills.
- Generating specialized Openclaw Skills from proven production code to enhance AI agent capabilities across multiple repos.
- Analyzing a repository to create a comprehensive project summary and technical debt map.
- Extracting customized Tailwind configurations or CSS variables into a portable, project-agnostic design system document.
How Pattern Extraction Works
- Discovery: The agent scans the root directory, configuration files (like package.json or tailwind.config), and documentation to map the tech stack and project structure.
- Categorization: Discovered elements are prioritized into categories such as Design Systems, UI Patterns, Architecture, and Workflows to ensure high-value data is processed first.
- Extraction: Valuable patterns are processed through specialized templates to generate specific outputs, including design system docs and project-agnostic Openclaw Skills.
- Validation: Each extracted item is checked against quality criteria to ensure it contains expert-level knowledge and adheres to strict formatting rules before being saved.
- Output: Finalized documentation and Openclaw Skills are written to local directories or staged for refinement across multiple repositories.
Pattern Extraction Setup
Install the extraction module using the OpenClaw hub CLI to begin analyzing your codebase:
npx clawhub@latest install extraction
Ensure you have access to the reference files like methodology-values.md and skill-quality-criteria.md within your local environment to guide the extraction logic and maintain the high standards expected of Openclaw Skills.
Pattern Extraction Data Schema & Taxonomy
The skill organizes extracted data into a specific directory structure to maintain clarity and reusability for Openclaw Skills:
| Directory | Content Type | Purpose |
|---|---|---|
docs/extracted/ |
Markdown Files | Stores project summaries, design system values, and architecture maps. |
ai/skills/ |
SKILL.md & References | Contains the modular Openclaw Skills derived from the codebase patterns. |
ai/staging/ |
Temporary Storage | Used when aggregating patterns from multiple projects before final refinement. |
Pattern Extraction Advanced Features
- Multi-project staging for consolidating patterns from various repositories into a unified standard for Openclaw Skills.
- Conflict detection to prevent duplicate Openclaw Skills by checking existing local skill libraries before creating new entries.
- Aesthetic direction capture which prioritizes the vibe and philosophy of a design system over raw code snippets.
- Automated validation against expert-knowledge thresholds to ensure high-quality agent interactions and accurate methodology recovery.
- Project-agnostic normalization that strips specific project names to ensure the resulting Openclaw Skills are portable.
SKILL.md
Loading
Related Openclaw Skills
Expo Native UI
A comprehensive toolkit for building production-ready native iOS and Android applications using Expo Router and modern React Native patterns.
Uniswap Token Swap Agent Skill
An autonomous DeFi skill that enables AI agents to execute, simulate, and validate token swaps across Uniswap V2, V3, V4, and UniswapX.
Estimation Patterns
A comprehensive framework for producing accurate, defensible software estimates using mathematical models and systematic decomposition.
E2E Testing Patterns
Build reliable, high-performance end-to-end test suites using industry-standard patterns for Playwright and Cypress.
Feature Specification
A technical meta-skill that converts persona documentation into detailed, implementable feature specifications with clear acceptance criteria.
Financial Data Visualization
A comprehensive set of patterns for building professional, dark-themed financial charts and trading user interfaces.
Featured*
