# ai-evals

Datasets, benchmarks, and evaluation metrics for AI agents. Name benchmark repositories `eval-<agent-name>-benchmark`.

This public repository is a top-level area of the [IT-One](https://git.it-one.pro/IT-One) organization.

## Collaboration

- Use short-lived branches and pull requests.
- Keep credentials out of Git; use organization-level Gitea Actions secrets.
- Add a focused README to every project repository created in this area.
- Use the [AI agent template](https://git.it-one.pro/IT-One/agent-template) for agent projects.