autoresearch

Automate iterative research loops over codebases with bounded iterations and logged metrics.

2|1|Updated Feb 27, 2026
One-click install
npx skills add https://github.com/spideynolove/claude-code-in-action --skill autoresearch-spideynolove
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: autoresearch
Source: https://github.com/spideynolove/claude-code-in-action/tree/main/25-autoresearch/.claude/skills/autoresearch
Command: npx skills add https://github.com/spideynolove/claude-code-in-action --skill autoresearch-spideynolove

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Autonomous, configurable research workflows for codebases, enabling teams to run goal-directed, iterative experiments that improve quality, reliability, and performance without manual chore work.

Core Features & Use Cases

  • End-to-end autoresearch loops: plan, execute, verify, and log iterations with mechanical metrics
  • Interactive setup with bounded Iterations and explicit scope to tailor runs to your project
  • Git-as-memory: leverage repository history to learn from prior experiments and avoid repeating failures
  • Knowledge-file generation (codebase-analysis.md, dependency-map.md, component-clusters.md) to align multiple agents
  • Chaining and downstream handoffs to debug, security, fix, ship, scenario, predict, learn

Quick Start

Start /autoresearch with your goal and scope, then let the system run indefinitely or for a fixed number of iterations.

Frequently Asked Questions about autoresearch

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

FAQPage Schema
How do I automate iterative research loops over a codebase?

Automated codebase research leverages repository history as memory to learn from prior experiments and avoid repeating failures. It generates knowledge files like codebase-analysis.md, dependency-map.md, and component-clusters.md to align multiple agents across the project.

How do I use git history to prevent repeated failures in code analysis?

Set up automated research by initiating the process with your specific goal and scope, then allowing the system to run indefinitely or for a fixed number of iterations. Every iteration logs reproducible metrics, guard checks, and a structured results trail for auditing.

Can I run goal-directed code analysis across multiple repositories?

Yes, you can chain automated research outputs to downstream handoffs for debugging, security, fixes, shipping, scenario testing, prediction, and learning. This allows the structured results trail to feed directly into subsequent workflow stages.

What are the limitations of autonomous codebase research loops?

Automated codebase research generates knowledge files including codebase-analysis.md, dependency-map.md, and component-clusters.md. These files capture the structured results trail to align multiple agents and provide an auditable record of each iteration.