alexsergivan/kwant
A task earns “Verified” status only after kwant runs the check itself and keeps the evidence.
kwant is a free, open-source app for teams of AI bots that share chats, hand off tasks and work on schedules. Version 0.9 adds a review step where a bot on another model reruns checks in a clean copy of the files; high-risk code also gets small mutations to test whether the tests catch them. An Inbox shows verdicts, evidence and anything that could not be checked, with options to approve, send back or roll back work. It runs on your computer or server using AI accounts or API keys you supply.
GitHub · LLM repos' note
kwant is a free, open-source app for teams of AI bots that share chats, hand off tasks and work on schedules. Version 0.9 adds a review step where a bot on another model reruns checks in a clean copy of the files; high-risk code also gets small mutations to test whether the tests catch them. An Inbox shows verdicts, evidence and anything that could not be checked, with options to approve, send back or roll back work. It runs on your computer or server using AI accounts or API keys you supply.
GitHub · LLM repos' note
score 4