[tarun_anand]

Congratulations!! Anchoring is important. What about other parts of the code review like coding guidelines, perf issues etc?

[dolftax]

We flag performance issues today alongside security and code quality. We're working on respecting AGENTS.md, detecting code complexity (AI generated code tends toward verbose, tangled logic), and letting users/teams define custom coding guidelines.

[tarun_anand]

The AI tools already have a rules engine for coding guidelines etc.

I guess the real question is can Deepsource be the "judge" of whether the guidelines were followed, NFR will be met by humans and AI alike