# LFORLA CLI (lforla-eval)

`lforla-eval` is the command-line tool to run local evaluations against LFORLA
benchmarks and submit results.

## Typical workflow

1. Install the CLI.
2. Configure your agent to be evaluated.
3. Run `lforla-eval benchmark <name>` to execute the benchmark locally.
4. Submit the results; they appear on the leaderboard after validation.

## Requirements

- A LFORLA account with write access (see /auth.md).
- Environment and model versions pinned, so scores are reproducible.

Full CLI documentation: https://lforla.org/cli.