AI Blocks › Build and ship software
Competing agent racer
Give it one task. Get several AI agents trying it in separate branches, scored, with the best one merged.
Competing agent racer is a free AI skill for the job "Build and ship software". Give the instructions to an AI assistant that supports the required files and tools, then review its output.
Free · No account · MIT licence · Made by Alireza Rezvani · Level: expert · Thinking: high
Made by Alireza Rezvani. See the original on GitHub; we copied it as it is, with its licence.
What it does
Several AI agents try the same task side by side, each in its own copy of your project, and the best result is merged. You stop betting on one attempt and pick from real alternatives with scores. You start a session with the task, how many agents to run and how to judge them. Each agent works in a separate git branch so they cannot clash. Ready-made patterns exist for speed tuning, cleaning up code, writing tests and fixing bugs. A coordinator ranks the results by a measured number or an AI judge, merges the winner and archives the rest. It needs git and a code project, and also works for comparing drafts of text.
You give it
- The task in one sentence
- A git project to work in
- How many agents to run
- A way to score results, such as a test or a timing
You get
- Several attempts in separate branches
- A ranked comparison
- The winning branch merged
- A status view of every agent
Use it when
- You have a slow function and want three different speed-ups to compare.
- A bug has several possible fixes and you want to see them tested.
- You need more test coverage and want the best set of tests.
- You want to compare content drafts and keep the strongest.
How to use it
- Add it to your AI once: download it and add the folder to Claude, ChatGPT or Codex as a skill.
- Ask for the job in your own words. Your AI opens the skill when it fits.
- It works best when your AI can run code, use git and start helper agents.