autoresearch

Automates end-to-end optimization loops for measurable metrics.

Updated Apr 6, 2026

Search Tags:#automation #git #optimization #metrics #journal #experiments #autoresearch

Ready to run

npx skills add https://github.com/Volinad/v-skills-public --skill autoresearch-volinad

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: autoresearch
Source: https://github.com/Volinad/v-skills-public/tree/main/autoresearch
Command: npx skills add https://github.com/Volinad/v-skills-public --skill autoresearch-volinad

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This skill automates end-to-end autonomous optimization loops for any measurable metric, enabling teams to systematically improve performance without manual trial-and-error.

Core Features & Use Cases

Supports two modes: parameter mode for numeric tuning and code mode for file edits and Git integration.
Runs hypothesize–experiment–analyze cycles, records a structured journal, and generates a final report for review.
Works across domains (trading strategies, code performance, prompts/configs) and can run overnight or for long-running sessions.

Quick Start

Start an autoresearch session to optimize a measurable metric by proposing hypotheses, running experiments, and reviewing the generated report.

Frequently Asked Questions about autoresearch

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

FAQPage Schema

How do I automate optimization loops for a measurable metric?▼

You can automate optimization loops by running hypothesize-experiment-analyze cycles that systematically improve a target metric. It supports parameter mode for numeric tuning and code mode for file edits with Git integration.

How does a structured optimization journal work for automated experiments?▼

A structured optimization journal records the results of automated hypothesize-experiment-analyze cycles. It logs parameter and code mode experiments across domains like trading strategies and code performance, concluding with a generated final report for review.

What do I need to run autonomous optimization experiments overnight?▼

Running autonomous optimization experiments overnight requires Python 3.10+, Git, a valid eval command, and a defined parameter space or program.md. These prerequisites enable long-running automated cycles for any measurable metric.

Can I use autonomous optimization for trading strategies and code performance?▼

Yes, autonomous optimization works across domains including trading strategies, code performance, prompts, and configurations. It supports both numeric parameter tuning and code modification modes to improve any measurable metric.

What is the best way to tune numeric parameters without manual trial-and-error?▼

The best way to tune parameters without manual trial-and-error is using parameter mode within an autonomous optimization loop. It automates the hypothesize-experiment-analyze cycle, systematically testing numeric values against your eval command to improve the target metric.

When should I avoid using automated optimization loops?▼

You should avoid automated optimization loops if your project lacks a valid eval command, a defined parameter space, or a program.md file. Without these prerequisites and a measurable metric, the autonomous hypothesize-experiment-analyze cycle cannot function properly.