Upstream watch — daily KB refresh (25/10/2026)
Our own summary of what moved upstream since the 24/10 run. Nothing is copied; each point cites its source.
Sources that changed
- make-all/tuya-local 7169307…c6ea798 (24–25/10/2026). Wi-Fi only, so these are leads for the Wi-Fi/LAN layer, not Zigbee couples:
- New product ids: Vibrix PureFlow30 purifier
igqljfdcj6hwjjir(DPs 1, 2, 3, 5, 6, 7, 8, 18; PR #6308, ozgurclcc); Pro Breeze PB-D-18W-WF dehumidifier006lnrrrsa2sayck(DPs 1–7, 11–13, 101; PR #6067, surfnterd); smart plug energy v3rhvvzh71kuirphg2(Wipro 16A) andjwv5jtbscfluha0g(10A); Aziot IR remotess6zyykmyh6xmfwzq,key54vrth5askhsj; Siguro purifier57rwsy9izugk0ts7; Simpal TY-WD thermostathqvhlveyqzrncz9w. - Parsing lesson (commit 0002e83, K-Rain KRX8 irrigation): a signal-strength value packed as the first byte of a long raw DP must be read as a signed int8 (RSSI is negative). Same rule applies to any raw Tuya block that carries RSSI.
- Behaviour fixes only (no new DP): Arlec fan 1 h timer option, Smartmi circulator preset/speed range, Vonroc heater preset case.
- Koenkk/zigbee-OTA b050897…df5e8e1 (25/10/2026 02:49 Europe/Paris): automatic download run that only touched the stack-info index and two vendor caches.
index.jsonis unchanged, so no new OTA image for our devices. - esphome/esphome b90e1e4 (PR #18525, bkbartk, merged 24/10/2026 01:52 Europe/Paris): Tuya fan "optimistic" mode. Idea for us: when a DP has never been reported, a write can fall back to the type implied by its function (speed → value, oscillation → bool) and the UI state can be updated optimistically until the device answers. Recorded in CROSS_APP_METHODS; not implemented (our fallback chain already re-reads after a write).
- zigpy/zha (6decce1, 24/10/2026): dependency bumps only (zigpy 2.4.0, bellows 1.2.1, zigpy-deconz 1.1.0). No quirk change.
- SmartThingsEdgeDrivers (3c3946e, 23/10/2026): Matter fingerprint only. Nothing for Zigbee/Tuya.
Sources checked and unchanged
zigbee-herdsman-converters (still eb823146c7, PR #13583 is a coordinator-detection fix, no Tuya device), zigbee-herdsman (c1dcb11), Zigbee2MQTT docs (2d9af8d), ZHA quirks (1d6598c), deCONZ DDF (a4c17ad), Blakadder (d96e8aa), tinytuya (3c0a687), localtuya (59c95cd), Johan's app code (328d24e), Athom zigbee-clusters (b859be7, 3.9.0), kkossev/Hubitat (pushed 20/10), ioBroker.zigbee (16/10), Domoticz (19/10), Domoticz-Zigbee stable9 and Tasmota Zigbee driver files (no commit since 24/10). The old Homey SDK docs repo (athombv/homey-apps-sdk-v3-docs) is gone; SDK facts are now taken from apps.developer.homey.app (links only).
Couples added (one reliable source each)
| couple | driver | source |
|---|---|---|
_TZ3000_pd9mpyh4 / TS0044 | scene_switch_4 | Device interview in JohanBendz/com.tuya.zigbee#1534 (motren, 24/10/2026): battery remote, 4 endpoints, EP1 clusters 0x0000/0x0001/0x0006/0xE000, EP2–4 genOnOff. Same layout as the zgyzgdua/wkai4ga5 TS0044 scene remotes already on that driver. |
Unknown-command harvest
No new unknown command, cluster or DP this run: no diagnostic mail since 24/10 (only Athom build notices), no new post in forum topics 140352 (last post #2267) and 26439 (#5537), no new issue on dlnraja/com.tuya.zigbee besides the digest, and the fetch-diags ledger entry of 25/10 was a GitHub notice with nothing to act on. No decoder was added because no source confirmed a new Zigbee semantic.