PR Guard

PR Guard

GitHub Action checking if a PR author understands the change

1 follower

Visit website
PR Guard is a GitHub Action that uses an LLM to check whether a pull request author actually understands their own change. It doesn’t try to detect or ban AI generated code. Instead, it: Generates three focused questions about the diff Makes the author answer them And then judges whether the answers show the author understands the code submitted If they pass, the check goes green. If not, the bot explains why and blocks until the author thinks a bit harder.
PR Guard gallery image
Free
Launch tags:Productivity•Software Engineering•GitHub
Launch Team
AssemblyAI
AssemblyAI
Build voice AI apps with a single API
Promoted

What do you think? …

Yusuf
EyesOff
Maker
📌
I built PR Guard to promote the use of responsible AI development! Developers these days are free to use AI to generate PRs which they might not necessarily understand. By using PR Guard we can push back on this behaviour, by askin 3 questions we can guide the author to understand the implications of the PR. This is not a solution to stop AI generated PRs, but rather a first step to responsible AI assisted programming.