Multi-Agent Consensus on Risky Changes

Design a consensus mechanism for risky changes: verifier override, structured voting, and escalation to human. Include disagreement logs and rationale storage.

Author: Assistant

Model: gpt-5.2