---
name: "ax-python-gepa"
description: "Use when writing Python code with `axllm` for GEPA, Pareto tradeoffs, reflection clients, metric budgets, optimizer state, and artifacts."
version: "22.0.3"
---
# Ax GEPA For Python

This skill helps an agent write Python code with the generated Ax package `axllm`. Use the generated package API, examples, and manifests; do not import TypeScript-only APIs unless you are editing the TypeScript package.

## When To Use

- Run the generated GEPA optimizer or inspect a GEPA artifact.
- Use BootstrapFewShot before GEPA when demonstrations should seed optimization.
- Track metric budgets, reflection calls, candidate state, and Pareto fronts.

## Package Facts

- Language: Python.
- Package: `axllm`.
- Package API docs: `API.md` and `axir-api.json`.
- Capability manifest: `axir-capabilities.json`.
- Runnable examples: `examples/`.
- Real network support: yes.
- Scripted no-key transport support: yes.
- Runtime profiles: `javascript-quickjs`, `python-pyodide`.

## Core Pattern

```python
from axllm import AxGEPA

engine = AxGEPA(reflection_client)
result = engine.optimize(request, evaluator)
```

## Relevant API Surface

- Optimizers: `optimize`, `AxBootstrapFewShot`, `AxGEPA`, `OptimizerEngine`, `OptimizerEvaluator`

## Guardrails

- Start from package examples for exact native syntax before inventing a new call shape.
- Use `provider-api` examples only when the user explicitly has provider credentials available.
- Use `no-key` examples for deterministic local checks and provider request mapping.
- Treat AxIR as the source of generated package truth: if package docs disagree with source code, update the compiler and regenerate packages.
- Do not copy repo-maintainer skills from `tools/*/skills/` into user packages.