Description

Previously, GEPA could only optimized system prompts that were defined in env.system_prompt. Now, we can optimize multiple prompt components, as well as dynamically loaded system prompts (like in rlm_env.py). We can also optimize env_groups as long as each environment exposes the same prompt components and starting prompts for those components. We can also optimize tool defs as well. The user can see which components are optimizable by vf-gepa <env> --see-components and specify which components to optimize with --components

Environments are now responsible for exposing prompt components. Default is exposing self.system_prompt and the tools registered in self.oai_tools.

wordle:

rlm_secrets:

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Test improvement

Testing

• All existing tests pass when running uv run pytest locally.
• New tests have been added to cover the changes

Checklist

• My code follows the style guidelines of this project as outlined in AGENTS.md
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published

Additional Notes