Plan Review Copilot
← back to flags
Audit trail

Reasoning trace

Clause: 3.2
Severity: medium
Engine: mock / mock-keyword-heuristic
Tokens: - in / - out
Latency: 0 ms
Called: 2026-07-22 03:57:12

Flag explanation

[SIMULATED - keyword heuristic, not real reasoning] This clause and the closest matching code section (11) both state numeric requirements, but the numbers don't match: this clause says 10, 6, the code section says 1011.8, 12, 3658. Worth a closer look to confirm whether this is a real conflict.

Prompt sent to model

Compare numbers in clause '3.2' vs candidate '11'

Raw model response

clause_numbers=['10', '6'] candidate_numbers=['1011.8', '12', '3658']