docs: follow the Home Assistant storage move to /volume1/home-assistant

AGENTS.md

@@ -34,8 +34,8 @@
 ## Live systems and safety
 
 - Live ESPHome configuration is on the NAS at
-  `/volume1/docker/homeassistant/esphome/`; HA configuration is at
-  `/volume1/docker/homeassistant/homeassistant/`.
+  `/volume1/home-assistant/esphome/config/`; HA configuration is at
+  `/volume1/home-assistant/homeassistant/config/`.
 - NAS access is read-only unless Florent explicitly authorizes a write, cache
   purge, compile, upload, or flash.
 - Validate exactly one canary before any fleet rollout. Never infer successful

CLAUDE.md

@@ -32,9 +32,9 @@ are in `docs/upstreaming-strategy.md` (`infra-3rr.26`); query GitHub or
   Do not infer the host type from a `/volume1/` path — that convention is not a
   vendor signature. Operational lessons (ACL/permission quirks, safe-write
   playbooks) live in the KB corpus `/home/flow/claude_project/knowledge-base/bundle/`.
-- ESPHome configs: `/volume1/docker/homeassistant/esphome/`
-- HA configs: `/volume1/docker/homeassistant/homeassistant/`
-- Plant label images: `/volume1/docker/homeassistant/esphome/plant_labels/`
+- ESPHome configs: `/volume1/home-assistant/esphome/config/`
+- HA configs: `/volume1/home-assistant/homeassistant/config/`
+- Plant label images: `/volume1/home-assistant/esphome/config/plant_labels/`
 - **Read-only by default** — write only on explicit user confirmation
 
 ## Device inventory

docs/pull-ota-eval.md

@@ -192,10 +192,10 @@ is legacy PlatformIO). Sketch of a post-build hook:
 ```bash
 # On the NAS, run after each Device Builder build of a pilot device
 DEV=<device>
-BIN="/volume1/docker/homeassistant/esphome/.esphome/build/$DEV/build/firmware.ota.bin"
+BIN="/volume1/home-assistant/esphome/config/.esphome/build/$DEV/build/firmware.ota.bin"
 WEB=/volume1/web/smartplant
 VER=$(yq -r '.esphome.project.version // .substitutions.project_version' \
-       /volume1/docker/homeassistant/esphome/$DEV.yaml)
+       /volume1/home-assistant/esphome/config/$DEV.yaml)
 MD5=$(md5sum "$BIN" | awk '{print $1}')
 cp "$BIN" "$WEB/firmware.ota.bin"
 jq -n --arg v "$VER" --arg md5 "$MD5" \

examples/multi-device/README.md

@@ -223,7 +223,7 @@ Standard fleet rollout after a merged package change:
    cancels in-flight Device Builder jobs and disarms queued updates, so run it
    before, never during, a batch. Before `reset`, confirm the pool still holds
    only the pinned receiver (read-only, prints no key material):
-   `ssh ugreen "sudo python3 -c 'import json;d=json.load(open(\"/volume1/docker/homeassistant/esphome/.offloader_pairings.json\"));print(\"local_in_pool\",d[\"include_local_in_pool\"]);[print(p[\"pin_sha256\"],p[\"status\"],p[\"enabled\"],p[\"label\"]) for p in d[\"pairings\"]]'"`
+   `ssh ugreen "sudo python3 -c 'import json;d=json.load(open(\"/volume1/home-assistant/esphome/config/.offloader_pairings.json\"));print(\"local_in_pool\",d[\"include_local_in_pool\"]);[print(p[\"pin_sha256\"],p[\"status\"],p[\"enabled\"],p[\"label\"]) for p in d[\"pairings\"]]'"`
    Run `$S --receiver-pin <pin> reset` for each additional approved peer.
 2. `$S status` — confirm which devices show `expected != deployed`.
 3. `$S update all` — waits up to 75 min for each device's hourly maintenance

scripts/esphome_fleet_update.py

@@ -59,7 +59,7 @@ async def main():
 
 
 def secret(name):
-    path = "/volume1/docker/homeassistant/esphome/secrets.yaml"
+    path = "/volume1/home-assistant/esphome/config/secrets.yaml"
     with open(path, encoding="utf-8") as handle:
         for line in handle:
             key, separator, value = line.partition(":")
@@ -84,7 +84,7 @@ def secret(name):
 
 
 def secret(name):
-    path = "/volume1/docker/homeassistant/esphome/secrets.yaml"
+    path = "/volume1/home-assistant/esphome/config/secrets.yaml"
     with open(path, encoding="utf-8") as handle:
         for line in handle:
             key, separator, value = line.partition(":")