ax-gepa

Name: ax-gepa
Availability: InStock
Author: jadecli

Generate AxGEPA optimization code for multi-objective agent-tree configurations.

1|1|Updated Mar 27, 2026

Search Tags:#optimization #agents #codegen #flows #pareto #multi-objective #ax-gepa

Ready to run

npx skills add https://github.com/jadecli/researchers --skill ax-gepa

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ax-gepa
Source: https://github.com/jadecli/researchers/tree/main/agentcrawls-ts/.claude/skills/ax-gepa
Command: npx skills add https://github.com/jadecli/researchers --skill ax-gepa

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Generates AxGEPA optimization code to efficiently optimize generators, flows, and agent trees using multi-objective criteria within the @ax-llm/ax framework.

Core Features & Use Cases

Automates creation of AxGEPA-based optimization pipelines for complex AI workflows including flows, agents, and trees.
Supports deterministic evaluation with compile(), Pareto-front exploration, and metrics-driven optimization.
Use Case: A developer wants to balance accuracy and latency across a distributed agent system.

Quick Start

Provide an AxGEPA-based optimization setup to optimize a given classifier or workflow.

Frequently Asked Questions about ax-gepa

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I optimize multi-objective agent trees using Pareto fronts?▼

Multi-objective agent tree optimization uses AxGEPA to explore Pareto fronts by evaluating deterministic metrics across configurations. You provide metrics, validationExamples, and maxMetricCalls to bound evaluation cost and return an optimized program.

What is multi-objective optimization for LLM workflows?▼

Multi-objective optimization for LLM workflows balances competing criteria like accuracy and latency. AxGEPA applies Pareto-front exploration to generators, flows, and agent trees, returning a set of optimal trade-off programs rather than a single solution.

How do I configure AxGEPA.compile for flow optimization?▼

To configure AxGEPA.compile for flow optimization, supply your metrics, validationExamples, and maxMetricCalls. This bounds the optimization cost and returns an optimized program alongside the evaluated Pareto front for your flow or agent tree.

Can I balance accuracy and latency across distributed agent systems?▼

Yes, you can balance accuracy and latency across distributed agent systems by applying AxGEPA multi-objective optimization. It evaluates your agent-tree configurations against multiple metrics to generate a Pareto front of optimal trade-offs.

Do I need validation examples to bound optimization costs in codegen?▼

Yes, validationExamples are required alongside metrics and maxMetricCalls to bound evaluation costs during AxGEPA optimization. These inputs ensure deterministic evaluation and scalable instruction optimization for your generators and flows.

Why does my multi-objective agent optimization return a Pareto front instead of one result?▼

Multi-objective agent optimization returns a Pareto front because it identifies multiple trade-off solutions across competing metrics rather than a single winner. AxGEPA.compile outputs the entire Pareto front alongside the optimized program.