Autoresearch

Generate skill variations, evaluate them, and commit the best to a pull request.

Updated Jul 16, 2026

Search Tags:#pull request #skill testing #code improvement #skill evolution #skill research

Ready to run

npx skills add https://github.com/enuno/noesis-aeon --skill autoresearch-enuno

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: Autoresearch
Source: https://github.com/enuno/noesis-aeon/tree/main/skills/autoresearch
Command: npx skills add https://github.com/enuno/noesis-aeon --skill autoresearch-enuno

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the process of evolving a skill by generating variations, evaluating them, and updating the best version, thus improving the performance and reliability of existing skills.

Core Features & Use Cases

Skill Evolution: Automatically generate variations of an existing skill to improve its performance.
Variation Generation: Create up to four distinct variations based on different theses.
Evaluation and Selection: Score each variation on various criteria and select the best one for implementation.
PR Creation: Commit the winning variation to a pull request for further review and implementation.

Quick Start

Evolve the skill 'token-movers' by running the command: autoresearch var=token-movers

Frequently Asked Questions about Autoresearch

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

FAQPage Schema

How do I automate code improvement for existing skills through variation testing?▼

Automate code improvement by generating up to four distinct skill variations, scoring them against predefined criteria, and committing the best performing version to a pull request for review.

What is the best way to evolve a skill automatically without manual code reviews for every change?▼

The best way to evolve a skill automatically is to run an automated research process that generates variations, evaluates them based on predefined criteria, and outputs a single winning pull request for final review.

How does variation generation work when testing skill performance improvements?▼

Variation generation works by creating up to four distinct versions of a skill based on different theses, evaluating their performance against criteria, and selecting the highest-scoring variation for implementation.

How do I submit improved code variations to a pull request for review?▼

Submit improved code variations by running the automated skill evolution command, which evaluates generated variations and commits the winning version directly to a pull request for review.

Can I use automated skill research to update existing code without manual evaluation?▼

Yes, you can use automated skill research to update existing code by reading context from memory, generating variations, and automatically evaluating them against predefined criteria to select the best version.

What are the limitations of automated skill evolution and variation testing?▼

Limitations of automated skill evolution include the need to read existing context from memory before generating variations, and the requirement to output changes as a pull request rather than direct implementation.