autoresearch

Name: autoresearch
Availability: InStock
Author: wjgoarxiv

Convert natural language research goals into iterative experimental research loops.

29|4|Updated Mar 15, 2026

Search Tags:#machine learning #hypothesis generation #experiment automation #autonomous research #iteration loop

Ready to run

npx skills add https://github.com/wjgoarxiv/autoresearch-skill --skill autoresearch-wjgoarxiv

💬 Prompt for ChatGPT / Claude Chat▼

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

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires Claude Code, Codex CLI, OpenCode, Gemini CLI, git, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill simplifies the autonomous research process by converting research goals into a self-driven loop that includes hypothesis generation, experimentation, and iteration without user intervention.

Core Features & Use Cases

Goal-driven Research Loop: Reads research.md files and automates research, including experiment evaluation.
Automated Hypothesis & Experimentation: Automatically suggests hypotheses based on past data and performs experiments.
Continuous Iteration: The system keeps running and improves based on feedback without further human input.

Quick Start

Run the research loop by providing a research.md file. For example: execute the "autoresearch" command without any subcommand and the agent will begin processing the provided research file.

Frequently Asked Questions about autoresearch

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

FAQPage Schema

How do I automate hypothesis generation and experiment execution from a natural language research goal?▼

Autonomous research automation converts natural language goals into a self-driven loop of hypothesis generation, experiment execution, and evaluation. By reading a research.md file, the system iterates without human intervention until the target metric is achieved.

What is an autonomous research agent and how does it handle continuous experimentation?▼

An autonomous research agent handles continuous experimentation by automatically suggesting hypotheses based on past data, running experiments, and evaluating results. It maintains a continuous iteration loop, improving based on feedback until the target metric is met.

How do I start an autonomous research loop using a markdown file?▼

To start an autonomous research loop, provide a research.md file containing your research goal. Execute the autoresearch command without subcommands, and the agent will begin processing the provided file to automate the experimental research stages.

Can I use Claude Code or Codex CLI to run an autonomous research iteration loop?▼

Yes, autonomous research loop execution supports integration with various tools including Claude Code, Codex CLI, OpenCode, and Gemini CLI. These dependencies facilitate the automated experimentation and hypothesis generation processes across different domains.

Does autonomous experiment automation work for machine learning tasks across different domains?▼

Yes, autonomous experiment automation is capable of handling tasks across various domains, including machine learning. It automates the research stages from understanding the goal to evaluating results and maintaining improvements without requiring further human input.

What are the limitations of autonomous research agents when evaluating experimental results?▼

Autonomous research agents rely on predefined target metrics to conclude the iteration loop. If the research goal in the research.md file is ambiguous or the target metric cannot be measured, the automated evaluation and continuous improvement process may not function correctly.