Journey event  ·  13 Sep 2026

A probe’s EC, pH or soil-temperature channel can freeze or rail without any alert: stuck detection covers moisture only, and the root-zone fault needs every pro

Found 2026-09-13 while scoring every injectable dev-bed fault in the LLM test loop's oracle. sensor_stuck / sensor_railed on pot{n}.ec_us, pot{n}.ph or pot{n}.soil_temp_c: ingest n

  state copied from the tracker, never improved

  event date 2026-09-13

Event date 2026-09-13
Version found
Version fixed
Expected kit firmware
Release file
SHA-256
Served count
Source https://app.notion.com/3da2b4cda37081e9b488fcea20653030

Found 2026-09-13 while scoring every injectable dev-bed fault in the LLM test loop’s oracle. sensor_stuck / sensor_railed on pot{n}.ec_us, pot{n}.ph or pot{n}.soil_temp_c: ingest nulls an out-of-range reading and records it under soil_implausible, and the moisture flatline drives probe{n}_sensor_stuck, but nothing alerts on a frozen or rejected EC/pH/soil-temp channel. binary_sensor.dsc_hub_root_zone_sensor_fault only fires when NO in-service probe has a plausible soil temperature, so one probe railing to −127 °C while the other reads is silent. Root steering and Need then work from a frozen or missing EC/pH. Suggested: per-probe held alerts for a rejected channel and a flatline on EC/pH/soil temp (spans tuned to each sensor’s resolution). The oracle treats these as a documented limit until then.

All events  ·  DSC  ·  CannaLib  ·  Site  ·  Firmware releases

Empty rows are honest

A blank cell means the tracker has no value for it. Nothing on this page is filled in to look complete.