EvalPal allows developers to trigger AI agent evaluation runs, track results, and list definitions via a streamlined API integration.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install evalpal
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 evalpal 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 EvalPal?
EvalPal is a specialized tool designed to bring rigorous AI agent testing into your development workflow. By integrating with Openclaw Skills, it enables developers to execute complex evaluation suites that measure episode quality, factual accuracy, and response consistency without leaving the terminal or chat environment.
This skill serves as a bridge between agent development and quality assurance, providing immediate feedback loops. It automates the heavy lifting of polling for results and formatting technical API responses into human-readable markdown reports, ensuring that your AI agents meet performance benchmarks consistently.
EvalPal Use Cases
- Validating AI agent performance after code changes or prompt adjustments.
- Automated quality assurance for customer support bots and factual accuracy checks.
- Monitoring long-running evaluation tasks with built-in status tracking.
- Discovering and listing available evaluation definitions across multiple AI projects.
How EvalPal Works
- The user initiates an evaluation request using the evalpal run command with a specific evaluation definition ID.
- The skill authenticates with the EvalPal API using the secure EVALPAL_API_KEY environment variable.
- A new evaluation run is triggered, and the skill enters a polling phase using exponential backoff to manage resource usage.
- Once the evaluation completes or reaches the 5-minute timeout, the skill fetches the raw data using curl and jq.
- The results are parsed and formatted into a structured markdown report, displaying pass/fail counts and specific test case details.
EvalPal Setup
To get started, ensure you have your API key from the EvalPal dashboard. Configure your environment as follows:
# Set your mandatory API key
export EVALPAL_API_KEY="sk_your_key_here"
# Optional: Override the default API URL if using a custom instance
export EVALPAL_API_URL="https://evalpal.dev"
Ensure that both curl and jq are installed on your system as they are required for API communication and data parsing.
EvalPal Data Schema & Taxonomy
The skill interacts with the EvalPal API and returns data structured for developer readability. The following table describes the primary data points returned during a run:
| Attribute | Description |
|---|---|
| Run ID | Unique identifier for the specific evaluation instance (e.g., run_abc123). |
| Status | Current state of the eval: running, passed, failed, or error. |
| Results | Breakdown of individual test cases (tc_001, tc_002) with success/failure indicators. |
| Metadata | Timestamps for start/end times and total duration in seconds. |
EvalPal Advanced Features
- Intelligent polling with exponential backoff to ensure reliability without hitting rate limits.
- Cross-project evaluation listing to manage complex AI workforce labs from a single interface.
- Granular error handling for HTTP status codes (401, 404, 429) and network timeouts.
- Secure credential management that ensures the EVALPAL_API_KEY is never exposed in logs or console output.
SKILL.md
Loading
Related Openclaw Skills
Client Outreach Automator
An AI-powered sales assistant that creates personalized, multi-touch outreach campaigns tailored to specific industries and business pain points.
GSD Orchestrator
An automation-first orchestration skill that executes GSD project management workflows through a headless CLI interface.
Model Handoff
The Model Handoff skill ensures seamless context continuity when switching between different LLM models by maintaining a persistent state file.
GitHub to OpenClaw Skill
A documentation-only reference guide for converting existing GitHub projects into modular Openclaw Skills.
HubSpot CRM Database Audit
A comprehensive diagnostic tool that audits HubSpot CRM portals across eight dimensions to identify data rot, deliverability issues, and automation inefficiencies.
Shopify Headless Commerce Strategy
A comprehensive evaluation and implementation framework for transitioning Shopify stores to a high-performance custom headless architecture.
Featured*
