Review bots failed in Projects

Hey guys,

I wanted to run review bots in a project but both review agents failed:

The security and Bugbot reviews could not complete. This environment does not expose the security-review or bugbot agent types, so both launches were rejected with:

Invalid enum value — expected generalPurpose, explore, computerUse, videoReview, cursor-guide, ci-investigator, best-of-n-runner, local-workflow, ai-architect, deployment-expert, performance-optimizer, or hostinger-deployment-reviewer.

A second attempt with the same prompt failed the same way. I did not substitute another agent or review the diff myself.

Could you please advise?

Thanks!

Hey @Daniel_Quisek, thanks for the report! This is how it works today rather than something broken: the /review commands only attach to cloud agents started from cursor.com/agents, the IDE, or the CLI, so agents a Project spins up as workers don’t get them, and retrying in the worker won’t change that.

To run the review now: open cursor.com/agents, start an agent on the branch the worker pushed, and send /review-bugbot (or /review-security). Your Project’s main chat should also work.