ForkFree

Agent skills and plugins

autoresearch

by uditgoenka

A skill pack for Claude Code and similar coding agents that runs a repeating loop: make one change, test it against a measurable score, keep it if it improves and undo it if not. It is for leaving an AI agent to improve code or other measurable work without constant supervision.

Proof level: ReadWorth a lookMIT

Get it on GitHubProject websiteCompare

No demo yet

Proof level: Read. We read its README and licence and ran our automatic checks. We have not run it yet. So this page is based on its README and our automatic checks.

The catch

Worth a lookEstablished, with a few things to check (listed below).

Check this first: Needs a paid Claude Code or other AI coding agent subscription or API usage, and unattended loops can use a lot of tokens.

Things to check

  • Needs a paid Claude Code or other AI coding agent subscription or API usage, and unattended loops can use a lot of tokens
  • Autonomous loops make and commit code changes on their own, so use it on a branch with backups
  • Hook guardrails are described as defence-in-depth, not a security sandbox
  • Some README claims, such as the research citations, should be checked before relying on them

Can you use autoresearch commercially?

Free to use, change and sell, including in paid products. Keep the copyright notice.

Not legal advice. Read the project's licence file before building a product on it.

What can you do with autoresearch?

  • Set a goal and a measurable score, then let the agent make and test changes automatically overnight
  • Use subcommands to plan work, find and fix bugs, run security reviews and generate documentation
  • Review a log of every experiment, with failed changes reverted automatically through git

How do you get autoresearch running?

It is installed as a skill into Claude Code, OpenCode, Codex, pi or Muse, so you need one of those agents already set up; the excerpt does not give exact install steps.

Follow the official instructions in the project README.

Make money with it

A freelance developer could offer overnight optimisation or bug-fixing for client codebases, using the loop to do the repetitive trial-and-error work.

An idea, not financial advice. Check the licence before charging for anything built on it.

Ways a business could use it

  • Improve a website's or app's test pass rate, speed or code quality overnight without paying a developer for each iteration
  • Run automated security and bug-hunting passes on a small product before release

Checked 11 Nov 2026. Summary drafted by AI from the project's README. Proof level 1 of 4, Read: we read its readme and licence and ran our automatic checks. we have not run it yet.

Similar tools in Agent skills and plugins

Worth a look

A skill for coding agents that runs a structured, multi-phase security audit of a codebase. It uses separate agents to find, challenge and verify vulnerabilities, then writes the results as machine-readable records and readable reports.

Proof level: Read 53.3k stars +1411 a day MIT
unslop heydamir
Check the licence

A skill for Claude Code and Codex that reviews a screen built with AI tools, finds the generic "AI look" such as purple gradients, glows and uneven spacing, and fixes it step by step. It is for making dashboards, apps and landing pages look designed rather than auto-generated.

Proof level: Read 18.9k stars +787 a day Unclear
agent_skills janestreet
New and promising

A small collection of agent skills, each a folder with a SKILL.md file of instructions for an AI agent. At present it holds one skill, a Socratic tutor that teaches a topic through questions, a skill map and spaced review.

Proof level: Read 10.3k stars +749 a day MIT