While it might sound unconventional to some users here, I believe Replit and (we as users in the long run) would benefit from a ‘Critical Agent’ Mode, similar to the recent release from Google’s Jules (ref: https://jules.google/docs/changelog/2026-01-26-1).
We often operate under the assumption that our logic is flawless and our app ideas are perfect. Replit Agent is excellent at executing exactly what we ask for, but this can sometimes lead to a ‘feedback vacuum’ where we build things that have fundamental flaws or weak PRDs (Product Requirement Documents).
The Feature Request: Give users the option to toggle on a ‘Critical’ mode. Instead of just saying ‘Yes’ and coding, the Agent would:
-
Challenge assumptions in the PRD.
-
Identify potential edge cases or UX friction points before writing code.
-
Offer ‘Devil’s Advocate’ feedback to help us avoid disappointment later in the development cycle.
It’s awesome that Replit Agent is so compliant, but having an optional ‘Reality Check’ mode would be a game-changer for serious builders.
Just my two cents 