Heater vent conflict fires on the firmware’s deliberate moisture-overflow venting: the brain flags heater on + exhaust ≥ 25 % with no RH-overflow exception, but

Found 2026-09-14 while reviewing a dev-bed LLM loop round (round 3 raised heater_vent_conflict on a healthy rig). That round’s direct cause was the simulator, which ignored the firmware’s heater no-extraction interlock (fixed on the dev branch, f484127). Checking the firmware showed a real brain gap behind it. Firmware: while heater demand is live, OUT drops to the 15 % fresh-air floor UNLESS rh_overflow (RH > stage RH max + 5), in which case OUT is at least 70 % by design, because that moisture must leave the building. The brain’s heater_vent_conflict is heat demand + heater relay on + outside exhaust ≥ 25 %, with no overflow exception, so it raises a cue on exactly the state the firmware chooses on purpose. Expected: exempt the overflow case (4×8 RH > RH max + 5), matching the firmware. The humidifier interlock has the same exemption, but a humidifier running while RH is over max + 5 is itself contradictory, so it is left as is.