Submit a tool for testing
Paste a GitHub URL for any MCP server, agent skill (OpenClaw, skills.sh), or SDK. We'll evaluate it across five dimensions.
Queued for evaluation.
We'll publish results when ready. Check back in about 30 minutes.
What happens after you submit
- 1. We clone your repository and detect the tool type (Node.js or Python).
- 2. The tool is built and started in an isolated Docker container with no network access.
- 3. Five test suites run: security, reliability, agent usability, compatibility, and code health.
- 4. Scores are aggregated into a weighted verdict (0-10) and published with full test evidence.