GitHub radar

Vercel’s Foreman: AI-Powered Agents That Automate GitHub Issue Resolution

Foreman is an open-source AI software factory template powered by the eve framework. It uses four autonomous AI agents to process each GitHub/Linear task: a Classifier, an Analyst, an Implementer, and a Reviewer. These agents work independently without any human intervention to create a draft pull request.

Foreman is an open source AI powered software factory template built using the eve framework. It supports integration with GitHub and Linear. All labelled tasks are processed by 4 agents — a Classifier which classifies requests and asks clarifying questions where necessary, an Analyst which creates an implementation plan using a live clone of the repository and sets acceptance criteria, an Implementer which executes the implementation plan in an isolated environment, runs relevant project tests and commits changes to a branch, and a Reviewer who reviews the diff (but not the thought process of the Implementer) and gives a verdict with an explanation. Even when Foreman isn’t deployed, there is a live factory memory which contains notes on the repository. This is shared amongst all the agents. Foreman can be deployed easily by pressing the Vercel button which will automate the setup of the GitHub & Linear integration.

Why a vibe-coder should care

Using Foreman, you can have your to-do lists managed in GitHub issues, while letting AI agents handle the implementation. You still have the last word on whether to merge or not, but you now have 4 agents who will do all the hard work of researching, planning, coding and testing for you, right from the description of the task till the draft pull request.

Open on GitHub