"Devllon team on IBM Bob 2.0 hackathon Hackathon"

Team Idea

An AI safety layer for coding agents that determines whether they understand enough of a codebase to safely make a requested change and knows when to stop and investigate instead. Our core team is complete.

Team Members

Muhammad Wajahat Jamshed Tawakkaly

0x-wajahat

Izaan Ahmad

izaanahmad

Aarez Absar

AarezSatti

AbdurRehman Danish

AbdurRehman_Danish

Submission

Noesis — Know When to Stop

Noesis is a safety layer for AI coding agents that investigates a codebase's hidden dependencies and history before allowing a change, stopping to gather evidence or ask a human instead of guessing.