An automated auditing and optimization tool designed to ensure the structural integrity and logical consistency of AI agent skills.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skill-check
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-check 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 Check?
Skill Check is a specialized utility focused on the static analysis and optimization of Openclaw Skills. It helps developers maintain high-quality agent instructions by identifying logic gaps, redundant descriptions, and structural deviations from standard protocols. The tool emphasizes the separation of concerns, ensuring that deterministic logic is moved into scripts while high-level workflows remain in the markdown documentation.
By using this skill, developers can transform loose instructions into professional-grade modules that are cross-compatible with various AI coding environments. It enforces the Freedom Matching Principle, which balances strict script-based execution for fragile tasks with flexible LLM reasoning for high-context decisions, ensuring your Openclaw Skills are both robust and adaptable.
Skill Check Use Cases
- Auditing existing AI agent skills for structural quality and logic gaps.
- Transitioning manual step-by-step instructions into automated Python scripts.
- Ensuring that Openclaw Skills remain tool-agnostic and free from hardcoded IDE dependencies.
- Generating standardized review reports to track the maturity of a skill's development.
How Skill Check Works
- The system initiates a structural analysis of the target skill directory using provided scripts.
- It identifies P0 and P1 issues based on structural dimensions like line count and asset placement.
- The LLM performs a deep logic audit to find contradictions, missing boundary conditions, or execution ambiguities.
- It checks for agent-specific coupling to ensure the skill works across different AI coding platforms.
- The tool generates an actionable optimization plan with direct steps to improve the skill's performance and reliability.
Skill Check Setup
To begin auditing your Openclaw Skills, ensure your skill directory contains at least a SKILL.md file. Use the following commands to run the analysis:
# Run a single audit on a specific skill directory
python3 scripts/analyze.py <skill-dir>
# Run an automated review-fix loop to iteratively improve the skill
python3 scripts/review-loop.py <skill-dir> --max-iterations 5
Skill Check Data Schema & Taxonomy
Skill Check enforces a standardized organization for Openclaw Skills to ensure discoverability and performance:
| Directory/File | Status | Description |
|---|---|---|
manifest.json |
Optional | Contains metadata like name, description, and category. |
SKILL.md |
Mandatory | The core workflow; must be concise and under 500 lines. |
scripts/ |
Recommended | Deterministic logic written in Python for high reliability. |
references/ |
Optional | External API docs, schemas, and long-form reference materials. |
assets/ |
Optional | Static resources such as images, templates, or sample data. |
Skill Check Advanced Features
- Iterative Self-Correction: Automatically detects and fixes structural errors through a recursive review loop.
- Tool-Agnostic Validation: Scans for and removes hardcoded references to specific IDEs or proprietary APIs.
- Priority-Based Reporting: Categorizes issues into P0 (critical) and P1 (recommended) for efficient debugging.
- Deterministic Extraction: Signals when complex reasoning should be converted into stable script-based logic.
SKILL.md
Loading
Related Openclaw Skills
Wukong
A framework to distill the thinking patterns of famous figures into structured six-layer cognitive models for expert-level problem analysis.
Stock Monitor
A comprehensive stock monitoring and technical analysis tool designed for A-share and Hong Kong market investors.
Daily Song Recommender
A dynamic AI music discovery skill that performs real-time web searches to provide personalized song and album recommendations.
Context Guardian
Context Guardian provides a robust framework for maintaining task continuity and preventing context loss in autonomous AI agents.
Decision Briefing Claw
An automated assistant for daily business data aggregation, KPI calculation, and multi-channel briefing delivery.
design-md: AI-Native Design Specifications
An automated system for creating and managing DESIGN.md files that allow AI agents to generate UI components strictly following your brand's design system.
Featured*
