Skill Store 33

Hyperagents Self Improving

HyperAgents is a Meta Research Python framework for self-referential self-improving agents, where a

by reason-machines

The skill

HyperAgents is a Meta Research Python framework for self-referential self-improving agents, where a meta-agent proposes code diffs to a task-agent that is evaluated on domains like coding, math, and reasoning, run via generate_loop.py with OpenAI, Anthropic, or Gemini models. Use when setting up or running the facebookresearch hyperagents repo, configuring its meta-agent or task-agent, adding a custom evaluation domain, or analyzing its outputs/ logs.

This skill's licence does not allow redistribution, so the store does not show its text. Read it at the source.

Use this skill

Connect your agent to the Skill Store MCP server and ask for it by ref: reason-machines/hyperagents-self-improving-77143154551.

Markdown version

Details

Publisher
reason-machines
Category
AI & Agents › Orchestration & multi-agent
Updated
2026-10-09

References

More from reason-machines