An automated auditing tool that evaluates the reliability and engineering quality of AI agent skills using the SRL framework.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skill-evaluator-srl
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-evaluator-srl 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 Evaluator SRL?
Skill Evaluator SRL is a specialized diagnostic tool designed to measure the Skill Reliability Level (SRL) of your agentic workflows. By combining AI-driven semantic analysis with deterministic Python script calculations, it provides a rigorous assessment of how dependable a skill truly is. This tool is essential for developers building Openclaw Skills who want to transition from experimental prompts to production-ready engineering.
The core philosophy of this tool is that a skill's reliability is proportional to its objective, externally verifiable facts versus its autonomous AI reasoning. By identifying engineering anchors and hallucination exposure points, Skill Evaluator SRL helps you build more robust Openclaw Skills that behave predictably in complex environments.
Skill Evaluator SRL Use Cases
- Pre-deployment quality audits for new Openclaw Skills to ensure high reliability.
- Automated batch testing of skill libraries to identify weak links in agentic logic.
- Continuous integration (CI) checks to prevent quality regression during skill updates.
- Detailed hallucination risk analysis for high-stakes AI automation tasks.
How Skill Evaluator SRL Works
- The evaluator identifies the target directory and validates the presence of the SKILL.md file.
- An AI agent reads the full skill content, including scripts and reference documents, to build a global logic model.
- The tool performs a multi-dimensional review, scoring categories like anchor density, traceability, and failure transparency.
- A dedicated Python script (srl_analyzer.py) performs weighted math, applies correction factors, and maps the results to an SRL grade.
- The system generates a comprehensive Markdown report and a structured JSON data file for integration.
Skill Evaluator SRL Setup
To begin evaluating your Openclaw Skills, ensure you have the evaluator skill installed in your project path. Run the analysis via the CLI:
# Standard evaluation via stdin
echo '<SKILL_JSON_DATA>' | python3 ./scripts/srl_analyzer.py
# Evaluate using an input file
python3 ./scripts/srl_analyzer.py --input /tmp/skill-data.json
# Generate a specific Markdown report
python3 ./scripts/srl_analyzer.py > .srl-report.md
Skill Evaluator SRL Data Schema & Taxonomy
The evaluator organizes its quality data into five primary dimensions. This structured approach ensures that Openclaw Skills are audited consistently:
| Dimension | Description | Weight |
|---|---|---|
| Engineering Anchor Density | Measures reliance on CLI, API, or DB calls. | 25% |
| Hallucination Exposure | Calculates the ratio of pure AI reasoning vs. verified steps. | 20% |
| Failure Transparency | Evaluates how clearly the skill reports errors. | 20% |
| Traceability | Checks if output data sources are citeable. | 20% |
| Reproducibility | Assesses state management and input-output stability. | 15% |
Skill Evaluator SRL Advanced Features
- Batch Assessment Mode: Audit dozens of Openclaw Skills simultaneously to generate a global ranking.
- Automated Quality Self-Checks: Built-in validation to ensure scoring logic remains consistent and evidence-based.
- Dynamic Correction Factors: Adjusts scores based on timeliness, robustness, and cascade stability to reflect real-world durability.
- Integrated Improvement Suggestions: Provides actionable, step-by-step instructions to fix specific architectural weaknesses.
SKILL.md
Loading
Related Openclaw Skills
UX Architect
A specialized tool for bridging design and development through technical UX architecture and robust CSS systems.
SEAP Shopping Skill
An intelligent shopping assistant for AI agents that automates product discovery, scheduled purchasing, and secure transaction execution.
accountant-priv-assistant
An AI-powered personal finance assistant that queries SQLite databases from Hapoalim, Isracard, and Max to provide deep financial insights.
Veteran Entrepreneur Video
This skill automates the creation of high-trust, fiduciary-aligned marketing videos for financial practices and wealth management firms.
Image Prompt Engineer
A technical prompt engineering framework designed to translate visual concepts into high-precision technical instructions for AI image generators.
ClawTrace
A diagnostic utility that analyzes OpenClaw logs and session context to generate detailed performance reports and optimization insights.
Featured*
