[hdmi_cec] CEC message decoder from the original branch

The decoder sources as they were before they were taken out of #9189 to keep the review small
(commit 11ded9dfbf on the author's branch). They turn a frame into a one-line text with the opcode
name and its operands, for the log. They are not wired into the merged bus core yet: the original
called the decoder from Frame::to_text() behind HDMI_CEC_USE_DECODER, with a decode_messages
option in the configuration.

.agents/skills/code-review/SKILL.md

@@ -0,0 +1,116 @@
+---
+name: code-review
+description: Review guidance for ESPHome pull requests. Use this when reviewing a pull request that changes ESPHome Python, C++, or component code, to check it against the project's coding conventions, embedded-systems memory rules, testing requirements, and breaking-change policy.
+---
+
+# Reviewing ESPHome pull requests
+
+ESPHome parses YAML into C++ firmware for memory-constrained microcontrollers
+(ESP32, ESP8266, RP2040, LibreTiny). Review changes with that in mind: RAM and
+flash are scarce, and code runs unattended for months.
+
+`AGENTS.md` in the repository root is the full contributor guide and the
+authority when it disagrees with this summary. The developer documentation at
+https://developers.esphome.io explains the component lifecycle and the reasoning
+behind these rules. This skill lists the concrete things worth flagging in a
+review; read `AGENTS.md` for the detail behind any item.
+
+Only raise findings that the diff actually introduces or changes. Do not ask for
+drive-by cleanup of pre-existing code the PR did not touch.
+
+## Memory and embedded constraints (highest value)
+
+Heap allocation after `setup()` is treated as a reliability bug, not a
+performance nit, because it fragments a small shared heap. Flag:
+
+- New heap allocation on a hot path or after setup that could be avoided.
+- `std::vector` where the size is known at compile time (use `std::array`, or
+  `StaticVector<T, N>` when a `push_back` API is needed) or fixed at runtime
+  init (use `FixedVector<T>`).
+- Listener / child-entity registration lists stored as `std::vector`; these have
+  a compile-time-known count and should use `cg.slot_counter()` plus
+  `StaticVector`.
+- `std::vector<uint8_t>` for a byte buffer that never grows: prefer
+  `std::unique_ptr<uint8_t[]>` or `std::array`.
+- `std::map` / `std::set` / `std::unordered_map` for small datasets (1-16
+  elements): a `std::vector` of a small struct with linear search is lighter.
+- `std::deque` anywhere: it allocates 512-byte blocks and should be avoided.
+- `std::string` storing a value set once from config: prefer `StringRef` (the
+  literal already lives in flash).
+- `std::string` / `std::to_string` / string-returning helpers on hot paths where
+  a buffer or view API exists.
+
+## C++ conventions
+
+- Include what you use: a file referencing a symbol must include the header that
+  declares it, even if it currently arrives transitively. New or changed uses of
+  a symbol need the matching include.
+- Prefix all member access with `this->`.
+- Naming: `lower_snake_case` for functions/methods/variables, `UpperCamelCase`
+  for classes/structs/enums, `UPPER_SNAKE_CASE` for namespace-scope constants,
+  trailing underscore on protected/private fields.
+- `enum class` values must be prefixed with the enum name in `UPPER_SNAKE_CASE`
+  (e.g. `UARTFlushResult::UART_FLUSH_RESULT_SUCCESS`). Bare names like `SUCCESS`,
+  `FAIL`, or `OK` collide with SDK macros on some platforms and break the build.
+- Prefer `const`/`enum` over `#define`; `#define` is only for conditional
+  compilation and code-generation sizes.
+- Never call `millis()` in a `loop()` body; use
+  `App.get_loop_component_start_time()`. A rate-limit gate below ~16 ms (the loop
+  period) does nothing.
+- Pick the timing primitive by cadence: gated `loop()` under 250 ms,
+  `set_interval` at 500 ms and above.
+- Do not override a base method to return the value it already returns (e.g.
+  `get_setup_priority()` returning `setup_priority::DATA`).
+- Wrap string literals passed as printf `%s` args in `LOG_STR_LITERAL()`.
+- Required, invariant dependencies should be constructor parameters, not setters.
+- Callback registration methods must be templated (`template<typename F>`), not
+  typed as `std::function`, so lightweight forwarders avoid a heap allocation.
+- Two-space indent, `using` over `typedef`, wrap at 120 columns.
+
+## Python conventions
+
+- Type-annotate every new function signature (params and return), new dataclass
+  fields, and new module-level variables. Import `ConfigType` from
+  `esphome.types`.
+- Use the walrus operator to avoid a double lookup, e.g.
+  `if (blah := config.get(CONF_BLAH)) is not None:`.
+- Reuse existing validators from `config_validation.py` (`cv.rename_key`,
+  `cv.has_exactly_one_key`, etc.) via `cv.All(...)` instead of hand-rolling.
+- `esphome/const.py` is frozen: no new `CONF_` constants there. Define them in
+  the component's own `.py`, or in `esphome/components/const/__init__.py` when
+  shared. The same constant defined in three or more component files fails CI.
+- State that must persist during code generation goes in `CORE.data` namespaced
+  under the component `DOMAIN` (a `@dataclass`), not module-level mutable globals.
+- Prefer callback-based triggers via `build_callback_automation()`; only use a
+  `Trigger<Ts...>` subclass when the forwarder needs mutable state.
+
+## Testing and coverage
+
+- New and changed lines and branches need test coverage, including defensive
+  early-returns, error paths, and no-op guards. A mocked-out function is not
+  covered; exercise the real call path too.
+- Component YAML tests live in `tests/components/<component>/`. Never define
+  buses (uart, i2c, spi, modbus) directly in a test file: pull them from
+  `tests/test_build_components/common/` through dict-style `packages:` so CI can
+  group builds. List-style packages or top-level merge keys block grouping.
+- Config-only checks use the `validate.*.yaml` prefix; compiled checks use
+  `test.*.yaml`.
+
+## Breaking changes and public API
+
+- Base classes under `esphome/core/` and documented config options are public
+  API. Undocumented `public` members of a component are internal.
+- A breaking change needs justification, a migration path in the PR description,
+  and a deprecation window where feasible (`ESPDEPRECATED` in C++,
+  `cv.rename_key(..., removed_in=...)` in Python). Changing a codegen-injected
+  lambda signature is not a breaking change.
+
+## Process and PR hygiene
+
+- PR titles start with a `[tag]` prefix: the component name (e.g. `[uart] ...`)
+  or `[core]` for shared code.
+- Prose in docs, comments, and commit messages should be plain English. Keep
+  inline comments short and only where the code is not self-explanatory; do not
+  restate what the code says.
+- Verify the PR fills out `.github/PULL_REQUEST_TEMPLATE.md` and adds
+  `CODEOWNERS` entries for a new component.

.clang-tidy.hash

@@ -1 +0,0 @@
-27aaab4e0ebfc10491720345aa746fc2dffa6a3985f73ec111b12dd99078d46f

.claude/settings.json

@@ -0,0 +1,16 @@
+{
+  "hooks": {
+    "SessionStart": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "d=\"${CLAUDE_PROJECT_DIR:-.}\"; [ -x \"$d/venv/bin/python\" ] || { mkdir -p \"$d/.temp\"; env -u VIRTUAL_ENV \"$d/script/setup\" >\"$d/.temp/setup.log\" 2>&1 || echo '{\"systemMessage\":\"script/setup failed; see .temp/setup.log\"}'; }",
+            "statusMessage": "Setting up dev environment (script/setup)...",
+            "timeout": 900
+          }
+        ]
+      }
+    ]
+  }
+}

.claude/skills

@@ -0,0 +1 @@
+../.agents/skills

.devcontainer/Dockerfile

@@ -1,4 +1,4 @@
-ARG BUILD_BASE_VERSION=2025.04.0
+ARG BUILD_BASE_VERSION=2026.06.1
 
 
 FROM ghcr.io/esphome/docker-base:debian-${BUILD_BASE_VERSION} AS base
@@ -31,7 +31,4 @@ RUN \
     platformio settings set enable_telemetry No \
     && platformio settings set check_platformio_interval 1000000
 
-COPY script/platformio_install_deps.py platformio.ini ./
-RUN ./platformio_install_deps.py platformio.ini --libraries --platforms --tools
-
 WORKDIR /workspaces

.devcontainer/devcontainer.json

@@ -15,7 +15,6 @@
     // uncomment and edit the path in order to pass through local USB serial to the container
     // , "--device=/dev/ttyACM0"
   ],
-  "appPort": 6052,
   // if you are using avahi in the host device, uncomment these to allow the
   // devcontainer to find devices via mdns
   //"mounts": [
@@ -41,7 +40,11 @@
       ],
       "settings": {
         "python.languageServer": "Pylance",
-        "python.pythonPath": "/usr/bin/python3",
+        // Use the container's pre-provisioned venv (built by the Dockerfile, outside the
+        // bind-mounted workspace) rather than a ./venv that may leak in from the host and
+        // mismatch the container's Python. See .devcontainer/Dockerfile (esphome-venv).
+        "python.defaultInterpreterPath": "/home/esphome/.local/esphome-venv/bin/python",
+        "python.terminal.activateEnvironment": true,
         "pylint.args": [
           "--rcfile=${workspaceFolder}/pyproject.toml"
         ],

.gitattributes

@@ -1,3 +1,5 @@
 # Normalize line endings to LF in the repository
 * text eol=lf
 *.png binary
+*.gif binary
+*.apng binary

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -9,6 +9,29 @@ body:
         If you have a feature request or enhancement, please [request them here instead][fr].
 
         [fr]: https://github.com/orgs/esphome/discussions
+  - type: markdown
+    attributes:
+      value: |
+        ## Use of AI in bug reports
+
+        AI tools are good at carrying out well-defined tasks, but they are not good at troubleshooting.
+        Please do NOT paste an AI-generated wall of text into the issue template - if the AI hasn't solved
+        your problem, its wild guesses are not likely to help.
+
+        Please DO include your own words and observations, compile/boot logs, and
+        especially a minimal reproducible example of your YAML configuration that demonstrates the problem.
+
+        It is however quite acceptable to use AI to translate your *own* report,
+        if you aren't a competent English speaker.
+
+        If you really think it will be useful to include an AI's analysis, preferably wrap it in a `<details>` block which will be collapsed by default.
+
+        If you are using AI to help solve a problem, rather than asking it to speculate about what the problem is,
+        it can be more useful to ask it to create a step-by-step troubleshooting procedure.
+        AI is also useful for generating boilerplate code, such as a minimal reproducible example of your YAML
+        configuration that demonstrates the problem.
+
+        Used properly, AI can be a useful tool to help you solve your problem, but don't let it get in the way.
   - type: textarea
     validations:
       required: true

.github/ISSUE_TEMPLATE/config.yml

@@ -8,7 +8,7 @@ contact_links:
     url: https://github.com/esphome/esphome-webserver/issues/new/choose
     about: Report an issue with the ESPHome web server.
   - name: Report an issue with the ESPHome Builder / Dashboard
-    url: https://github.com/esphome/dashboard/issues/new/choose
+    url: https://github.com/esphome/device-builder/issues/new/choose
     about: Report an issue with the ESPHome Builder / Dashboard.
   - name: Report an issue with the ESPHome API client
     url: https://github.com/esphome/aioesphomeapi/issues/new/choose

.github/PULL_REQUEST_TEMPLATE.md

@@ -6,6 +6,7 @@
 
 - [ ] Bugfix (non-breaking change which fixes an issue)
 - [ ] New feature (non-breaking change which adds functionality)
+- [ ] New developer-facing feature (adds functionality for component developers; no end-user configuration change)
 - [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected) — [policy](https://developers.esphome.io/contributing/code/#what-constitutes-a-c-breaking-change)
 - [ ] Developer breaking change (an API change that could break external components) — [policy](https://developers.esphome.io/contributing/code/#what-is-considered-public-c-api)
 - [ ] Undocumented C++ API change (removal or change of undocumented public methods that lambda users may depend on) — [policy](https://developers.esphome.io/contributing/code/#c-user-expectations)
@@ -20,6 +21,10 @@
 
 - esphome/esphome.io#<esphome.io PR number goes here>
 
+**Pull request in [developers.esphome.io](https://github.com/esphome/developers.esphome.io) with developer documentation (if applicable):**
+
+- esphome/developers.esphome.io#<developers.esphome.io PR number goes here>
+
 ## Test Environment
 
 - [ ] ESP32

.github/actions/build-image/action.yaml

@@ -15,11 +15,6 @@ inputs:
     description: "Version to build"
     required: true
     example: "2023.12.0"
-  base_os:
-    description: "Base OS to use"
-    required: false
-    default: "debian"
-    example: "debian"
 runs:
   using: "composite"
   steps:
@@ -47,7 +42,7 @@ runs:
 
     - name: Build and push to ghcr by digest
       id: build-ghcr
-      uses: docker/build-push-action@f9f3042f7e2789586610d6e8b85c8f03e5195baf # v7.2.0
+      uses: docker/build-push-action@c3c9e263c25d99ce0380d002d59b67737d91b0dc # v7.4.0
       env:
         DOCKER_BUILD_SUMMARY: false
         DOCKER_BUILD_RECORD_UPLOAD: false
@@ -60,7 +55,6 @@ runs:
         build-args: |
           BUILD_TYPE=${{ inputs.build_type }}
           BUILD_VERSION=${{ inputs.version }}
-          BUILD_OS=${{ inputs.base_os }}
         outputs: |
           type=image,name=ghcr.io/${{ steps.tags.outputs.image_name }},push-by-digest=true,name-canonical=true,push=true
 
@@ -73,7 +67,7 @@ runs:
 
     - name: Build and push to dockerhub by digest
       id: build-dockerhub
-      uses: docker/build-push-action@f9f3042f7e2789586610d6e8b85c8f03e5195baf # v7.2.0
+      uses: docker/build-push-action@c3c9e263c25d99ce0380d002d59b67737d91b0dc # v7.4.0
       env:
         DOCKER_BUILD_SUMMARY: false
         DOCKER_BUILD_RECORD_UPLOAD: false
@@ -86,7 +80,6 @@ runs:
         build-args: |
           BUILD_TYPE=${{ inputs.build_type }}
           BUILD_VERSION=${{ inputs.version }}
-          BUILD_OS=${{ inputs.base_os }}
         outputs: |
           type=image,name=docker.io/${{ steps.tags.outputs.image_name }},push-by-digest=true,name-canonical=true,push=true
 

.github/actions/cache-arduino8266/action.yml

@@ -0,0 +1,39 @@
+name: Cache Arduino ESP8266
+description: >
+  Resolve the pinned Arduino core and xtensa toolchain versions and cache the
+  native ESP8266 install (~110 MB framework + toolchain; no ccache store, the
+  seed job saves before any compile runs). Exports
+  ESPHOME_ARDUINO8266_PREFIX to the job so every later step installs into
+  the cached path; the Python venv must already be restored. Mirrors
+  cache-esp-idf: only dev-branch pushes write the shared cache, everything
+  else restores.
+runs:
+  using: composite
+  steps:
+    - name: Resolve the native toolchain cache key
+      # Versions are pinned in code, not a hashable file; resolve them so a
+      # bump changes the cache key. Assignment form so errexit catches a
+      # resolver failure.
+      id: version
+      shell: bash
+      run: |
+        # One owner for the install prefix: exported here and referenced by
+        # the cache steps below via env, so the caller's install and the
+        # cached path cannot diverge.
+        echo "ESPHOME_ARDUINO8266_PREFIX=$HOME/.esphome-arduino8266" >> "$GITHUB_ENV"
+        . venv/bin/activate
+        key=$(python -c 'from esphome.components.esp8266 import RECOMMENDED_ARDUINO_FRAMEWORK_VERSION as f; from esphome.arduino8266.framework import FRAMEWORK_RELEASES, TOOLCHAIN_VERSION as t; print(f"{FRAMEWORK_RELEASES[f].tag}-{t}")')
+        [ -n "$key" ] || exit 1
+        echo "key=$key" >> "$GITHUB_OUTPUT"
+    - name: Cache the native toolchain (write on dev)
+      if: github.ref == 'refs/heads/dev'
+      uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
+      with:
+        path: ${{ env.ESPHOME_ARDUINO8266_PREFIX }}
+        key: ${{ runner.os }}-esp8266-native-${{ steps.version.outputs.key }}
+    - name: Restore the native toolchain (off dev)
+      if: github.ref != 'refs/heads/dev'
+      uses: actions/cache/restore@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
+      with:
+        path: ${{ env.ESPHOME_ARDUINO8266_PREFIX }}
+        key: ${{ runner.os }}-esp8266-native-${{ steps.version.outputs.key }}

.github/actions/cache-clang-tidy-idedata/action.yml

@@ -0,0 +1,50 @@
+name: Cache clang-tidy idedata
+description: >
+  Cache the clang-tidy idedata and the headers it references under .temp
+  (headers only, about 30MB per env). Run after restore-python and cache-esp-idf.
+inputs:
+  environment:
+    description: 'clang-tidy environment (e.g. esp32-idf-tidy).'
+    required: true
+runs:
+  using: composite
+  steps:
+    - name: Compute cache key
+      id: key
+      shell: bash
+      run: |
+        . venv/bin/activate
+        [ -n "${{ inputs.environment }}" ] || { echo "::error::cache-clang-tidy-idedata: 'environment' input is empty"; exit 1; }
+        hash=$(python -c 'import sys; sys.path.insert(0, "script"); from clang_tidy_hash import idedata_cache_hash; print(idedata_cache_hash("${{ inputs.environment }}"))')
+        pyver=$(python -c 'import platform; print(platform.python_version())')
+        # Generating idedata is what installs ESP-IDF; never skip it over a missing
+        # install. This also skips the save, so a dev run that installs ESP-IDF
+        # warms the idedata cache on the next run.
+        if [ -d ~/.esphome-idf/frameworks ]; then
+          echo "skip=false" >> "$GITHUB_OUTPUT"
+        else
+          echo "ESP-IDF install missing, not using the clang-tidy idedata cache"
+          echo "skip=true" >> "$GITHUB_OUTPUT"
+        fi
+        echo "key=${{ runner.os }}-tidy-idedata-${{ inputs.environment }}-$hash-py$pyver" >> "$GITHUB_OUTPUT"
+        {
+          echo "path<<EOF"
+          printf '%s\n' '.temp/idedata-*.json' '.temp/idedata-*.hash'
+          printf '.temp/**/*.%s\n' h hpp hh hxx inc inl ipp tpp
+          echo "EOF"
+        } >> "$GITHUB_OUTPUT"
+    # Mirror cache-esp-idf: write on dev, restore-only on PRs. The post-step
+    # save only runs when the job succeeded, so a failed generation is never saved.
+    # Extend the extension list if a component ships extensionless headers.
+    - name: Cache clang-tidy idedata (write on dev)
+      if: (github.ref == 'refs/heads/dev' || contains(github.event.pull_request.labels.*.name, 'ci-cache-write')) && steps.key.outputs.skip != 'true'
+      uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
+      with:
+        path: ${{ steps.key.outputs.path }}
+        key: ${{ steps.key.outputs.key }}
+    - name: Cache clang-tidy idedata (restore-only off dev)
+      if: github.ref != 'refs/heads/dev' && !contains(github.event.pull_request.labels.*.name, 'ci-cache-write') && steps.key.outputs.skip != 'true'
+      uses: actions/cache/restore@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
+      with:
+        path: ${{ steps.key.outputs.path }}
+        key: ${{ steps.key.outputs.key }}

.github/actions/cache-esp-idf/action.yml

@@ -0,0 +1,70 @@
+name: Cache ESP-IDF
+description: >
+  Resolve the pinned ESP-IDF version and cache the native ESP-IDF install
+  (toolchains + source) at ~/.esphome-idf. Every job that installs ESP-IDF
+  natively (clang-tidy for IDF/Arduino and the component test batches) shares
+  one cache, since the install is identical: ESPHOME_IDF_DEFAULT_TARGETS
+  defaults to "all", and _get_configured_targets() in espidf/toolchain.py
+  skips per-variant narrowing whenever CI is set, so all toolchains are
+  present regardless of the chip a job builds.
+  Callers must set env ESPHOME_ESP_IDF_PREFIX: ~/.esphome-idf and have the
+  Python venv already restored.
+inputs:
+  framework:
+    description: 'Which pinned IDF version to key on: "espidf" (recommended) or "arduino".'
+    default: espidf
+  restore-only:
+    description: >
+      When "true", only restore -- never save the cache, even on dev. Use from
+      jobs that may not produce an ESP-IDF install (e.g. a component batch with
+      no esp32 target), so a partial/empty install is never written to the key.
+    default: "false"
+runs:
+  using: composite
+  steps:
+    - name: Resolve ESP-IDF version for cache key
+      # The native-IDF version is pinned in code, not in any file that feeds the
+      # other cache keys, so resolve it explicitly. Keying on it means the cache
+      # invalidates on a version bump (actions/cache never overwrites a key).
+      # Also key on the Python version: the cached IDF venv links to the
+      # runner's toolcache interpreter and is reinstalled every run after a
+      # runner image bump.
+      id: version
+      shell: bash
+      run: |
+        . venv/bin/activate
+        if [ "${{ inputs.framework }}" = "arduino" ]; then
+          version=$(python -c 'from esphome.components.esp32 import ARDUINO_FRAMEWORK_VERSION_LOOKUP as A, ARDUINO_IDF_VERSION_LOOKUP as L; print(L[A["recommended"]])')
+        else
+          version=$(python -c 'from esphome.components.esp32 import ESP_IDF_FRAMEWORK_VERSION_LOOKUP as L; print(L["recommended"])')
+        fi
+        echo "version=$version" >> "$GITHUB_OUTPUT"
+        echo "python-version=$(python -c 'import platform; print(platform.python_version())')" >> "$GITHUB_OUTPUT"
+    # Mirror the adjacent PlatformIO cache: only dev-branch runs write the
+    # shared cache (so it lives in the default-branch scope readable by all
+    # PRs), and PRs are restore-only -- they never push multi-GB artifacts into
+    # their own scope / the repo quota (e.g. on a version-bump PR). The
+    # ci-cache-write label lets a PR write into its own scope to test the hit path;
+    # that costs about 1GB of the repo cache quota per run, so remove it when done.
+    # -slim: bump when prune-esp-idf changes what it removes; a key is never overwritten.
+    - name: Cache ESP-IDF install (write on dev)
+      if: (github.ref == 'refs/heads/dev' || contains(github.event.pull_request.labels.*.name, 'ci-cache-write')) && inputs.restore-only != 'true'
+      uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+      with:
+        path: ~/.esphome-idf
+        key: ${{ runner.os }}-esphome-idf-${{ steps.version.outputs.version }}-py${{ steps.version.outputs.python-version }}-slim
+    - name: Cache ESP-IDF install (restore-only off dev)
+      if: github.ref != 'refs/heads/dev' && !contains(github.event.pull_request.labels.*.name, 'ci-cache-write') || inputs.restore-only == 'true'
+      uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+      with:
+        path: ~/.esphome-idf
+        key: ${{ runner.os }}-esphome-idf-${{ steps.version.outputs.version }}-py${{ steps.version.outputs.python-version }}-slim
+    # Install explicitly so the prune below sees the toolchains on a cache miss
+    # too, instead of the install happening inside the first build step.
+    - name: Install ESP-IDF
+      shell: bash
+      run: |
+        . venv/bin/activate
+        python -c 'from esphome.espidf.framework import check_esp_idf_install; check_esp_idf_install("${{ steps.version.outputs.version }}")'
+    - name: Prune ESP-IDF install
+      uses: ./.github/actions/prune-esp-idf

.github/actions/cache-sdk-nrf/action.yml

@@ -0,0 +1,49 @@
+name: Cache nRF Connect SDK
+description: >
+  Resolve the pinned sdk-nrf version and cache the native sdk-nrf install
+  (west workspace, Zephyr SDK toolchain, python env) at ~/.esphome-sdk-nrf.
+  Every job that installs sdk-nrf natively (the nrf52 clang-tidy job and,
+  once the component tests build natively, their batches) shares one cache.
+  Callers must set env ESPHOME_SDK_NRF_PREFIX: ~/.esphome-sdk-nrf and have
+  the Python venv already restored.
+inputs:
+  restore-only:
+    description: >
+      When "true", only restore -- never save the cache, even on dev. Use from
+      jobs that may not produce a complete install (e.g. a component batch
+      that fails mid-install), so a partial install is never written.
+    default: "false"
+runs:
+  using: composite
+  steps:
+    - name: Resolve sdk-nrf and toolchain versions for cache key
+      # Both versions are pinned in code, not in any file that feeds the
+      # other cache keys, so resolve them explicitly. Keying on them means
+      # the cache invalidates when either is bumped (actions/cache never
+      # overwrites a key).
+      id: version
+      shell: bash
+      run: |
+        . venv/bin/activate
+        version=$(python -c '
+        from esphome.components.nrf52 import RECOMMENDED_SDK_NRF_VERSION
+        from esphome.components.nrf52.framework import TOOLCHAIN_VERSION
+        print(f"{RECOMMENDED_SDK_NRF_VERSION}-{TOOLCHAIN_VERSION}")')
+        echo "version=$version" >> "$GITHUB_OUTPUT"
+    # Mirror cache-esp-idf: only dev-branch runs write the shared cache (so it
+    # lives in the default-branch scope readable by all PRs); PRs are
+    # restore-only and never push multi-GB artifacts into their own scope.
+    - name: Cache nRF Connect SDK install (write on dev)
+      if: github.ref == 'refs/heads/dev' && inputs.restore-only != 'true'
+      uses: actions/cache@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
+      with:
+        path: ~/.esphome-sdk-nrf
+        # yamllint disable-line rule:line-length
+        key: ${{ runner.os }}-esphome-sdk-nrf-${{ steps.version.outputs.version }}-${{ hashFiles('esphome/components/nrf52/requirements.txt') }}
+    - name: Cache nRF Connect SDK install (restore-only off dev)
+      if: github.ref != 'refs/heads/dev' || inputs.restore-only == 'true'
+      uses: actions/cache/restore@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
+      with:
+        path: ~/.esphome-sdk-nrf
+        # yamllint disable-line rule:line-length
+        key: ${{ runner.os }}-esphome-sdk-nrf-${{ steps.version.outputs.version }}-${{ hashFiles('esphome/components/nrf52/requirements.txt') }}

.github/actions/prune-esp-idf/action.yml

@@ -0,0 +1,34 @@
+name: Prune ESP-IDF install
+description: >
+  Remove the picolibc sysroots (1.1GB of the 3.9GB install) from the native
+  ESP-IDF toolchains; IDF 5.x links newlib. Skipped when an IDF 6 install is
+  present, which links picolibc (see esp32/__init__.py).
+runs:
+  using: composite
+  steps:
+    - name: Prune picolibc
+      shell: bash
+      run: |
+        shopt -s nullglob
+        prefix="${ESPHOME_ESP_IDF_PREFIX:-$HOME/.esphome-idf}"
+        prefix="${prefix/#\~/$HOME}"
+        for fw in "$prefix"/frameworks/*/; do
+          case "$(basename "$fw")" in
+            [6-9].*) echo "IDF $(basename "$fw") installed, keeping picolibc"; exit 0 ;;
+          esac
+        done
+        n=0
+        for dir in "$prefix"/tools/*-esp-elf/*/*-esp-elf/picolibc; do
+          echo "Removing $dir ($(du -sh "$dir" | cut -f1))"
+          rm -rf "$dir"
+          n=$((n + 1))
+        done
+        # The marker rides along in the cache entry so a restored slim tree stays quiet.
+        if [ "$n" -gt 0 ]; then
+          touch "$prefix/.picolibc-pruned"
+        elif [ -d "$prefix/tools" ] && [ ! -f "$prefix/.picolibc-pruned" ]; then
+          echo "::warning::no picolibc sysroots matched under $prefix/tools"
+        fi
+        if [ -d "$prefix" ]; then
+          du -sh "$prefix"
+        fi

.github/actions/restore-python/action.yml

@@ -17,12 +17,12 @@ runs:
   steps:
     - name: Set up Python ${{ inputs.python-version }}
       id: python
-      uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+      uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
       with:
         python-version: ${{ inputs.python-version }}
     - name: Restore Python virtual environment
       id: cache-venv
-      uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+      uses: actions/cache/restore@55cc8345863c7cc4c66a329aec7e433d2d1c52a9 # v6.1.0
       with:
         path: venv
         # yamllint disable-line rule:line-length
@@ -32,9 +32,12 @@ runs:
       # detects the activated venv via ``VIRTUAL_ENV`` so the venv layout
       # downstream jobs rely on is preserved.
       if: steps.cache-venv.outputs.cache-hit != 'true'
-      uses: astral-sh/setup-uv@08807647e7069bb48b6ef5acd8ec9567f424441b # v8.1.0
+      uses: astral-sh/setup-uv@1c37ad07a6a961277cf70c0d37d6f313000f5884 # v10.3.0
       with:
         enable-cache: true
+        # Pull request saves land in per-PR scopes nothing else can
+        # reuse; dev pushes seed the shared copy instead.
+        save-cache: ${{ github.event_name != 'pull_request' }}
         # Pin uv version so the action does not have to fetch the
         # manifest from raw.githubusercontent.com on every cache
         # miss; that fetch flakes on Windows runners.
@@ -46,7 +49,7 @@ runs:
         python -m venv venv
         source venv/bin/activate
         python --version
-        uv pip install -r requirements.txt -r requirements_test.txt
+        uv pip install -r requirements.txt -r requirements_dev.txt -r requirements_test.txt
         uv pip install -e .
     - name: Create Python virtual environment
       if: steps.cache-venv.outputs.cache-hit != 'true' && runner.os == 'Windows'
@@ -55,5 +58,5 @@ runs:
         python -m venv venv
         source ./venv/Scripts/activate
         python --version
-        uv pip install -r requirements.txt -r requirements_test.txt
+        uv pip install -r requirements.txt -r requirements_dev.txt -r requirements_test.txt
         uv pip install -e .

.github/scripts/auto-label-pr/constants.js

@@ -13,6 +13,7 @@ module.exports = {
     'merging-to-release',
     'merging-to-beta',
     'chained-pr',
+    'stacked-pr',
     'core',
     'small-pr',
     'medium-pr',
@@ -22,11 +23,13 @@ module.exports = {
     'has-tests',
     'needs-tests',
     'needs-docs',
+    'needs-developer-docs',
     'needs-codeowners',
     'too-big',
     'labeller-recheck',
     'bugfix',
     'new-feature',
+    'new-feature-developer',
     'breaking-change',
     'developer-breaking-change',
     'undocumented-api-change',
@@ -40,5 +43,17 @@ module.exports = {
     // Keep matching the old esphome-docs name during the transition period
     /https:\/\/github\.com\/esphome\/esphome-docs\/pull\/\d+/,
     /esphome\/esphome-docs#\d+/
+  ],
+
+  DEVELOPER_DOCS_PR_PATTERNS: [
+    /https:\/\/github\.com\/esphome\/developers\.esphome\.io\/pull\/\d+/,
+    /esphome\/developers\.esphome\.io#\d+/
+  ],
+
+  // Files whose developer-facing changes are documented via Python docstrings
+  // only - developers.esphome.io has no reference page for them yet, so PRs
+  // touching nothing but these files (and tests/) skip needs-developer-docs.
+  DEV_DOCS_EXEMPT_FILES: [
+    'esphome/config_validation.py'
   ]
 };

.github/scripts/auto-label-pr/detectors.js

@@ -1,4 +1,4 @@
-const { DOCS_PR_PATTERNS } = require('./constants');
+const { DOCS_PR_PATTERNS, DEVELOPER_DOCS_PR_PATTERNS, DEV_DOCS_EXEMPT_FILES } = require('./constants');
 const {
   COMPONENT_REGEX,
   detectComponents,
@@ -33,16 +33,54 @@ async function fetchPrFileContent(github, context, path) {
   }
 }
 
+// Check whether a pull request is part of a GitHub stack.
+//
+// GitHub's stacked pull request feature adds a `stack` object to the pull
+// request resource. It is present on every pull request in the stack -
+// including the bottom one, whose base is already `dev` - and is absent
+// entirely on standalone pull requests.
+//
+// The `pull_request_target` webhook payload is not guaranteed to carry this
+// field, so fall back to asking the API when it is missing. Guessing wrong
+// here is costly: a stacked pull request mistaken for a manually chained one
+// gets a label that blocks merging.
+async function isStackedPr(github, context) {
+  const pr = context.payload.pull_request;
+  if (pr.stack != null) {
+    return true;
+  }
+
+  try {
+    const { owner, repo } = context.repo;
+    const { data } = await github.rest.pulls.get({
+      owner,
+      repo,
+      pull_number: pr.number,
+    });
+    return data.stack != null;
+  } catch (error) {
+    // Treat an API failure as "not stacked" so a chained pull request still
+    // gets its blocking label rather than silently slipping through.
+    console.log('Failed to check stack membership:', error.message);
+    return false;
+  }
+}
+
 // Strategy: Merge branch detection
-async function detectMergeBranch(context) {
+async function detectMergeBranch(github, context) {
   const labels = new Set();
   const baseRef = context.payload.pull_request.base.ref;
+  const defaultBranch = context.payload.repository.default_branch;
 
   if (baseRef === 'release') {
     labels.add('merging-to-release');
   } else if (baseRef === 'beta') {
     labels.add('merging-to-beta');
-  } else if (baseRef !== 'dev') {
+  } else if (await isStackedPr(github, context)) {
+    // GitHub manages the merge order for a stack, so these are not blocked.
+    labels.add('stacked-pr');
+  } else if (baseRef !== defaultBranch) {
+    // A chain built by hand: it must not merge until its base branch does.
     labels.add('chained-pr');
   }
 
@@ -147,19 +185,9 @@ async function detectCoreChanges(changedFiles) {
 }
 
 // Strategy: PR size detection
-async function detectPRSize(prFiles, totalAdditions, totalDeletions, totalChanges, isMegaPR, SMALL_PR_THRESHOLD, MEDIUM_PR_THRESHOLD, TOO_BIG_THRESHOLD) {
+async function detectPRSize(prFiles, totalAdditions, totalDeletions, isMegaPR, SMALL_PR_THRESHOLD, MEDIUM_PR_THRESHOLD, TOO_BIG_THRESHOLD) {
   const labels = new Set();
 
-  if (totalChanges <= SMALL_PR_THRESHOLD) {
-    labels.add('small-pr');
-    return labels;
-  }
-
-  if (totalChanges <= MEDIUM_PR_THRESHOLD) {
-    labels.add('medium-pr');
-    return labels;
-  }
-
   const testAdditions = prFiles
     .filter(file => file.filename.startsWith('tests/'))
     .reduce((sum, file) => sum + (file.additions || 0), 0);
@@ -167,7 +195,24 @@ async function detectPRSize(prFiles, totalAdditions, totalDeletions, totalChange
     .filter(file => file.filename.startsWith('tests/'))
     .reduce((sum, file) => sum + (file.deletions || 0), 0);
 
-  const nonTestChanges = (totalAdditions - testAdditions) - (totalDeletions - testDeletions);
+  const nonTestAdditions = totalAdditions - testAdditions;
+  const nonTestDeletions = totalDeletions - testDeletions;
+
+  // small/medium count churn (additions + deletions) so a balanced refactor isn't undersized.
+  const nonTestChurn = nonTestAdditions + nonTestDeletions;
+
+  if (nonTestChurn <= SMALL_PR_THRESHOLD) {
+    labels.add('small-pr');
+    return labels;
+  }
+
+  if (nonTestChurn <= MEDIUM_PR_THRESHOLD) {
+    labels.add('medium-pr');
+    return labels;
+  }
+
+  // too-big uses net line delta (additions - deletions), matching the review message in reviews.js.
+  const nonTestChanges = nonTestAdditions - nonTestDeletions;
 
   // Don't add too-big if mega-pr label is already present
   if (nonTestChanges > TOO_BIG_THRESHOLD && !isMegaPR) {
@@ -238,6 +283,7 @@ async function detectPRTemplateCheckboxes(context) {
   const checkboxPatterns = [
     { pattern: /- \[x\] Bugfix \(non-breaking change which fixes an issue\)/i, label: 'bugfix' },
     { pattern: /- \[x\] New feature \(non-breaking change which adds functionality\)/i, label: 'new-feature' },
+    { pattern: /- \[x\] New developer-facing feature \(adds functionality for component developers; no end-user configuration change\)/i, label: 'new-feature-developer' },
     { pattern: /- \[x\] Breaking change \(fix or feature that would cause existing functionality to not work as expected\)/i, label: 'breaking-change' },
     { pattern: /- \[x\] Developer breaking change \(an API change that could break external components\)/i, label: 'developer-breaking-change' },
     { pattern: /- \[x\] Undocumented C\+\+ API change \(removal or change of undocumented public methods that lambda users may depend on\)/i, label: 'undocumented-api-change' },
@@ -348,12 +394,14 @@ async function detectRequirements(allLabels, prFiles, context, hasYamlLoadable)
   const labels = new Set();
 
   // Check for missing tests
-  if ((allLabels.has('new-component') || allLabels.has('new-platform') || allLabels.has('new-feature')) && !allLabels.has('has-tests')) {
+  if ((allLabels.has('new-component') || allLabels.has('new-platform') || allLabels.has('new-feature') || allLabels.has('new-feature-developer')) && !allLabels.has('has-tests')) {
     labels.add('needs-tests');
   }
 
   // Check for missing docs.
-  // `new-feature` (PR-body checkbox) always counts. `new-component` / `new-platform`
+  // `new-feature` (PR-body checkbox) always counts. `new-feature-developer` is
+  // deliberately excluded here: its docs live on developers.esphome.io and are
+  // checked separately below. `new-component` / `new-platform`
   // only count when at least one newly added file defines a top-level CONFIG_SCHEMA,
   // i.e. the new component/platform is actually loadable from YAML.
   const docsEligible =
@@ -369,6 +417,22 @@ async function detectRequirements(allLabels, prFiles, context, hasYamlLoadable)
     }
   }
 
+  // Check for missing developer docs. `new-feature-developer` requires a
+  // developers.esphome.io PR link, unless every changed file outside tests/ is
+  // in DEV_DOCS_EXEMPT_FILES (core validators documented via docstrings only).
+  if (allLabels.has('new-feature-developer')) {
+    const prBody = context.payload.pull_request.body || '';
+    const nonTestFiles = prFiles
+      .map(file => file.filename)
+      .filter(file => !file.startsWith('tests/'));
+    const onlyExemptFiles = nonTestFiles.every(file => DEV_DOCS_EXEMPT_FILES.includes(file));
+    const hasDevDocsLink = DEVELOPER_DOCS_PR_PATTERNS.some(pattern => pattern.test(prBody));
+
+    if (!onlyExemptFiles && !hasDevDocsLink) {
+      labels.add('needs-developer-docs');
+    }
+  }
+
   // Check for missing CODEOWNERS
   if (allLabels.has('new-component')) {
     const codeownersModified = prFiles.some(file =>

.github/scripts/auto-label-pr/index.js

@@ -53,7 +53,7 @@ module.exports = async ({ github, context }) => {
   });
   const currentLabels = currentLabelsData.map(label => label.name);
   const managedLabels = currentLabels.filter(label =>
-    label.startsWith('component: ') || MANAGED_LABELS.includes(label)
+    label.startsWith('component: ') || label.startsWith('platform: ') || MANAGED_LABELS.includes(label)
   );
 
   // Check for mega-PR early - if present, skip most automatic labeling
@@ -88,7 +88,7 @@ module.exports = async ({ github, context }) => {
 
   // Early exit for release and beta branches only
   if (baseRef === 'release' || baseRef === 'beta') {
-    const branchLabels = await detectMergeBranch(context);
+    const branchLabels = await detectMergeBranch(github, context);
     const finalLabels = Array.from(branchLabels);
 
     console.log('Computed labels (merge branch only):', finalLabels.join(', '));
@@ -118,12 +118,12 @@ module.exports = async ({ github, context }) => {
     deprecatedResult,
     maintainerAccess
   ] = await Promise.all([
-    detectMergeBranch(context),
+    detectMergeBranch(github, context),
     detectComponentPlatforms(changedFiles, apiData),
     detectNewComponents(github, context, prFiles),
     detectNewPlatforms(github, context, prFiles, apiData),
     detectCoreChanges(changedFiles),
-    detectPRSize(prFiles, totalAdditions, totalDeletions, totalChanges, isMegaPR, SMALL_PR_THRESHOLD, MEDIUM_PR_THRESHOLD, TOO_BIG_THRESHOLD),
+    detectPRSize(prFiles, totalAdditions, totalDeletions, isMegaPR, SMALL_PR_THRESHOLD, MEDIUM_PR_THRESHOLD, TOO_BIG_THRESHOLD),
     detectDashboardChanges(changedFiles),
     detectGitHubActionsChanges(changedFiles),
     detectCodeOwner(github, context, changedFiles),

.github/scripts/auto-label-pr/reviews.js

@@ -42,8 +42,8 @@ function generateReviewMessages(finalLabels, originalLabelCount, deprecatedInfo,
     let message = `${TOO_BIG_MARKER}\n### 📦 Pull Request Size\n\n`;
 
     message +=
-      `Hey @${prAuthor}, thanks for the contribution! Just a heads up, ` +
-      `this PR is on the large side `;
+      `Hey @${prAuthor}, thanks for the contribution. This PR is on the ` +
+      `large side `;
 
     if (tooManyLabels && tooManyChanges) {
       message +=
@@ -57,16 +57,31 @@ function generateReviewMessages(finalLabels, originalLabelCount, deprecatedInfo,
     }
 
     message += `, which makes it harder for maintainers to review.\n\n`;
+    let condition;
+    if (tooManyLabels && tooManyChanges) {
+      condition =
+        `is at most ${TOO_BIG_THRESHOLD} lines (excluding tests) and touches ` +
+        `at most ${MAX_LABELS} components/areas`;
+    } else if (tooManyLabels) {
+      condition = `touches at most ${MAX_LABELS} components/areas`;
+    } else {
+      condition = `is at most ${TOO_BIG_THRESHOLD} lines (excluding tests)`;
+    }
     message +=
-      `Smaller, focused PRs tend to be reviewed much faster since they ` +
-      `fit into the short gaps between other maintainer work; large ones ` +
-      `often have to wait for a rare long uninterrupted block of time. ` +
-      `If you can break this up into smaller pieces that can be reviewed ` +
-      `independently, it will almost certainly land faster overall.\n\n`;
+      `**This PR will not be reviewed until it ${condition}.** We know ` +
+      `that is not what you want to hear after this much work, and we want ` +
+      `to be upfront so you do not wait for a review that is not coming: ` +
+      `please split the change. The label is removed automatically once ` +
+      `the PR ${condition}.\n\n`;
     message +=
-      `Before putting more time in, it's also worth popping into ` +
-      `\`#devs\` on [Discord](https://esphome.io/chat) so we can help ` +
-      `you scope things and flag anything already in flight.\n\n`;
+      `There is no exception you can request. Rarely, a maintainer agrees in ` +
+      `\`#devs\` on [Discord](https://esphome.io/chat) to review a change that ` +
+      `genuinely cannot be split as one piece and adds the \`mega-pr\` label, ` +
+      `but that is rare enough that you should plan on splitting rather than ` +
+      `wait for it. Smaller, focused PRs are picked up much faster because ` +
+      `they fit into the short gaps between other maintainer work. For a new ` +
+      `component, the first PR brings the hub and a single platform (for ` +
+      `example \`sensor\`); each further platform is its own PR on top of it.\n\n`;
     message +=
       `For more details (including how to split the work up), see: ` +
       `https://developers.esphome.io/contributing/submitting-your-work/` +
@@ -214,6 +229,7 @@ async function handleMaintainerAccessComment(github, context, maintainerAccess)
 }
 
 module.exports = {
+  generateReviewMessages,
   handleReviews,
   handleMaintainerAccessComment
 };

.github/scripts/auto-label-pr/tests/detectors.test.js

@@ -1,6 +1,14 @@
 const { describe, it } = require('node:test');
 const assert = require('node:assert/strict');
-const { detectNewPlatforms, detectNewComponents } = require('../detectors');
+const {
+  detectMergeBranch,
+  detectNewPlatforms,
+  detectNewComponents,
+  detectPRSize,
+  detectPRTemplateCheckboxes,
+  detectRequirements,
+} = require('../detectors');
+const { MANAGED_LABELS } = require('../constants');
 
 // Minimal GitHub API mock — only repos.getContent is called by detectNewPlatforms/detectNewComponents
 // to check for CONFIG_SCHEMA in newly added files.
@@ -29,6 +37,122 @@ const API_DATA = {
 const WITH_SCHEMA = 'CONFIG_SCHEMA = cv.Schema({})';
 const WITHOUT_SCHEMA = 'CODEOWNERS = ["@esphome/core"]';
 
+// ---------------------------------------------------------------------------
+// detectMergeBranch
+// ---------------------------------------------------------------------------
+
+// Builds a fresh context for detectMergeBranch tests instead of mutating the
+// shared CONTEXT fixture above (which other describe blocks rely on).
+function makeMergeContext(baseRef, { stack, defaultBranch = 'dev' } = {}) {
+  const pull_request = { number: 1, base: { ref: baseRef } };
+  if (stack !== undefined) {
+    pull_request.stack = stack;
+  }
+  return {
+    repo: { owner: 'esphome', repo: 'esphome' },
+    payload: { pull_request, repository: { default_branch: defaultBranch } }
+  };
+}
+
+// A GitHub API mock exposing only rest.pulls.get, with a call counter so
+// tests can assert whether the API fallback was actually invoked.
+function makeStackGithub({ stack = null, error = null } = {}) {
+  const state = { calls: 0 };
+  const github = {
+    rest: {
+      pulls: {
+        get: async () => {
+          state.calls++;
+          if (error) throw error;
+          return { data: { stack } };
+        }
+      }
+    }
+  };
+  return { github, state };
+}
+
+const STACK_INFO = { base: { ref: 'dev' }, id: 71540, number: 17978, position: 3, size: 3 };
+
+describe('detectMergeBranch', () => {
+  it('base ref release adds merging-to-release only and never checks the stack', async () => {
+    const { github, state } = makeStackGithub({ stack: STACK_INFO });
+    const context = makeMergeContext('release', { stack: STACK_INFO });
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['merging-to-release']);
+    assert.equal(state.calls, 0);
+  });
+
+  it('base ref beta adds merging-to-beta only and never checks the stack', async () => {
+    const { github, state } = makeStackGithub({ stack: STACK_INFO });
+    const context = makeMergeContext('beta', { stack: STACK_INFO });
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['merging-to-beta']);
+    assert.equal(state.calls, 0);
+  });
+
+  it('stack present on the webhook payload adds stacked-pr without calling the API', async () => {
+    const { github, state } = makeStackGithub();
+    const context = makeMergeContext('feature-branch', { stack: STACK_INFO });
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['stacked-pr']);
+    assert.equal(state.calls, 0);
+  });
+
+  it('stack absent from payload falls back to the API and adds stacked-pr', async () => {
+    const { github, state } = makeStackGithub({ stack: STACK_INFO });
+    const context = makeMergeContext('feature-branch');
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['stacked-pr']);
+    assert.equal(state.calls, 1);
+  });
+
+  it('bottom of a stack (base ref dev, stack present) still adds stacked-pr', async () => {
+    const { github, state } = makeStackGithub();
+    const context = makeMergeContext('dev', { stack: STACK_INFO });
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['stacked-pr']);
+    assert.equal(state.calls, 0);
+  });
+
+  it('not stacked, base ref not dev adds chained-pr', async () => {
+    const { github } = makeStackGithub({ stack: null });
+    const context = makeMergeContext('feature-branch');
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['chained-pr']);
+  });
+
+  it('not stacked, base ref dev adds no labels', async () => {
+    const { github } = makeStackGithub({ stack: null });
+    const context = makeMergeContext('dev');
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), []);
+  });
+
+  it('a failed stack lookup falls back to not-stacked, so a feature-branch base adds chained-pr', async () => {
+    const { github, state } = makeStackGithub({ error: new Error('API unavailable') });
+    const context = makeMergeContext('feature-branch');
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['chained-pr']);
+    assert.equal(state.calls, 1);
+  });
+
+  it('base ref matches default branch adds no labels', async () => {
+    const { github } = makeStackGithub({ stack: null });
+    const context = makeMergeContext('other', { defaultBranch: 'other' });
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), []);
+  });
+
+  it('base ref dev when the default branch is main adds chained-pr', async () => {
+    const { github } = makeStackGithub({ stack: null });
+    const context = makeMergeContext('dev', { defaultBranch: 'main' });
+    const labels = await detectMergeBranch(github, context);
+    assert.deepEqual(Array.from(labels).sort(), ['chained-pr']);
+  });
+
+});
+
 // ---------------------------------------------------------------------------
 // detectNewPlatforms
 // ---------------------------------------------------------------------------
@@ -145,3 +269,198 @@ describe('detectNewComponents', () => {
     assert.equal(result.labels.size, 0);
   });
 });
+
+// ---------------------------------------------------------------------------
+// detectPRTemplateCheckboxes
+// ---------------------------------------------------------------------------
+
+const NEW_FEATURE_LINE = '- [x] New feature (non-breaking change which adds functionality)';
+const DEV_FEATURE_LINE = '- [x] New developer-facing feature (adds functionality for component developers; no end-user configuration change)';
+const DEV_FEATURE_LINE_UNTICKED = '- [ ] New developer-facing feature (adds functionality for component developers; no end-user configuration change)';
+
+function makeBodyContext(body) {
+  return { payload: { pull_request: { body } } };
+}
+
+describe('detectPRTemplateCheckboxes', () => {
+  it('ticked developer-facing feature checkbox adds new-feature-developer only', async () => {
+    const labels = await detectPRTemplateCheckboxes(makeBodyContext(DEV_FEATURE_LINE));
+    assert.ok(labels.has('new-feature-developer'));
+    assert.ok(!labels.has('new-feature'));
+  });
+
+  it('unticked developer-facing feature checkbox adds no label', async () => {
+    const labels = await detectPRTemplateCheckboxes(makeBodyContext(DEV_FEATURE_LINE_UNTICKED));
+    assert.ok(!labels.has('new-feature-developer'));
+  });
+
+  it('ticked new feature checkbox does not add new-feature-developer', async () => {
+    const labels = await detectPRTemplateCheckboxes(makeBodyContext(NEW_FEATURE_LINE));
+    assert.ok(labels.has('new-feature'));
+    assert.ok(!labels.has('new-feature-developer'));
+  });
+});
+
+// ---------------------------------------------------------------------------
+// detectRequirements
+// ---------------------------------------------------------------------------
+
+describe('detectRequirements', () => {
+  // PR body without any docs-PR link.
+  const NO_DOCS_CONTEXT = makeBodyContext('Just a description, no docs link.');
+  const USER_DOCS_CONTEXT = makeBodyContext('Docs: esphome/esphome.io#1234');
+  const DEV_DOCS_CONTEXT = makeBodyContext('Docs: esphome/developers.esphome.io#1234');
+  const DEV_DOCS_URL_CONTEXT = makeBodyContext('Docs: https://github.com/esphome/developers.esphome.io/pull/1234');
+
+  // File sets: a normal source change vs. one confined to the exempt core validators.
+  const SOURCE_FILES = [
+    { filename: 'esphome/components/foo/foo.py' },
+    { filename: 'tests/components/foo/common.yaml' },
+  ];
+  const VALIDATOR_FILES = [
+    { filename: 'esphome/config_validation.py' },
+    { filename: 'tests/unit_tests/test_config_validation.py' },
+  ];
+
+  it('new-feature-developer without has-tests adds needs-tests but not needs-docs', async () => {
+    const labels = await detectRequirements(new Set(['new-feature-developer']), SOURCE_FILES, NO_DOCS_CONTEXT, false);
+    assert.ok(labels.has('needs-tests'));
+    assert.ok(!labels.has('needs-docs'));
+  });
+
+  it('new-feature-developer with has-tests does not add needs-tests', async () => {
+    const labels = await detectRequirements(new Set(['new-feature-developer', 'has-tests']), SOURCE_FILES, NO_DOCS_CONTEXT, false);
+    assert.ok(!labels.has('needs-tests'));
+  });
+
+  it('new-feature without a docs link still adds needs-docs', async () => {
+    const labels = await detectRequirements(new Set(['new-feature', 'has-tests']), [], NO_DOCS_CONTEXT, false);
+    assert.ok(labels.has('needs-docs'));
+  });
+
+  it('new-feature-developer without a developer docs link adds needs-developer-docs', async () => {
+    const labels = await detectRequirements(new Set(['new-feature-developer', 'has-tests']), SOURCE_FILES, NO_DOCS_CONTEXT, false);
+    assert.ok(labels.has('needs-developer-docs'));
+  });
+
+  it('a developers.esphome.io shorthand link satisfies needs-developer-docs', async () => {
+    const labels = await detectRequirements(new Set(['new-feature-developer', 'has-tests']), SOURCE_FILES, DEV_DOCS_CONTEXT, false);
+    assert.ok(!labels.has('needs-developer-docs'));
+  });
+
+  it('a developers.esphome.io URL link satisfies needs-developer-docs', async () => {
+    const labels = await detectRequirements(new Set(['new-feature-developer', 'has-tests']), SOURCE_FILES, DEV_DOCS_URL_CONTEXT, false);
+    assert.ok(!labels.has('needs-developer-docs'));
+  });
+
+  it('a user docs (esphome.io) link does not satisfy needs-developer-docs', async () => {
+    const labels = await detectRequirements(new Set(['new-feature-developer', 'has-tests']), SOURCE_FILES, USER_DOCS_CONTEXT, false);
+    assert.ok(labels.has('needs-developer-docs'));
+  });
+
+  it('a developer docs link does not satisfy needs-docs for new-feature', async () => {
+    const labels = await detectRequirements(new Set(['new-feature', 'has-tests']), [], DEV_DOCS_CONTEXT, false);
+    assert.ok(labels.has('needs-docs'));
+  });
+
+  it('changes confined to core validator files are exempt from needs-developer-docs', async () => {
+    const labels = await detectRequirements(new Set(['new-feature-developer', 'has-tests']), VALIDATOR_FILES, NO_DOCS_CONTEXT, false);
+    assert.ok(!labels.has('needs-developer-docs'));
+  });
+
+  it('validator changes mixed with other source files are not exempt', async () => {
+    const prFiles = [...VALIDATOR_FILES, { filename: 'esphome/components/foo/foo.py' }];
+    const labels = await detectRequirements(new Set(['new-feature-developer', 'has-tests']), prFiles, NO_DOCS_CONTEXT, false);
+    assert.ok(labels.has('needs-developer-docs'));
+  });
+});
+
+// ---------------------------------------------------------------------------
+// MANAGED_LABELS
+// ---------------------------------------------------------------------------
+
+describe('MANAGED_LABELS', () => {
+  it('includes new-feature-developer so the workflow syncs it', () => {
+    assert.ok(MANAGED_LABELS.includes('new-feature-developer'));
+  });
+
+  it('includes needs-developer-docs so the workflow syncs it', () => {
+    assert.ok(MANAGED_LABELS.includes('needs-developer-docs'));
+  });
+});
+
+// ---------------------------------------------------------------------------
+// detectPRSize
+// ---------------------------------------------------------------------------
+
+describe('detectPRSize', () => {
+  const SMALL = 30;
+  const MEDIUM = 100;
+  const TOO_BIG = 1000;
+
+  function size(prFiles, isMegaPR = false) {
+    const totalAdditions = prFiles.reduce((sum, file) => sum + (file.additions || 0), 0);
+    const totalDeletions = prFiles.reduce((sum, file) => sum + (file.deletions || 0), 0);
+    return detectPRSize(prFiles, totalAdditions, totalDeletions, isMegaPR, SMALL, MEDIUM, TOO_BIG);
+  }
+
+  it('counts only non-test changes toward small-pr', async () => {
+    // 10 source + 5000 test lines -> non-test churn of 10 is still small.
+    const labels = await size([
+      { filename: 'esphome/components/foo/foo.cpp', additions: 10, deletions: 0 },
+      { filename: 'tests/components/foo/test.esp32-idf.yaml', additions: 5000, deletions: 0 },
+    ]);
+    assert.ok(labels.has('small-pr'));
+    assert.equal(labels.size, 1);
+  });
+
+  it('counts additions and deletions as churn (not net delta)', async () => {
+    // A balanced refactor (40 added, 40 removed) is 80 lines of churn -> medium, not small.
+    const labels = await size([
+      { filename: 'esphome/components/foo/foo.cpp', additions: 40, deletions: 40 },
+    ]);
+    assert.ok(labels.has('medium-pr'));
+    assert.equal(labels.size, 1);
+  });
+
+  it('labels medium-pr when non-test changes exceed small threshold', async () => {
+    const labels = await size([
+      { filename: 'esphome/components/foo/foo.cpp', additions: 60, deletions: 0 },
+      { filename: 'tests/components/foo/test.esp32-idf.yaml', additions: 5000, deletions: 0 },
+    ]);
+    assert.ok(labels.has('medium-pr'));
+    assert.equal(labels.size, 1);
+  });
+
+  it('uses net delta (not churn) for too-big', async () => {
+    // 600 added + 600 removed: 1200 churn (above too-big) but 0 net delta -> not too-big.
+    const labels = await size([
+      { filename: 'esphome/components/foo/foo.cpp', additions: 600, deletions: 600 },
+    ]);
+    assert.equal(labels.size, 0);
+  });
+
+  it('labels too-big when non-test changes exceed the big threshold', async () => {
+    const labels = await size([
+      { filename: 'esphome/components/foo/foo.cpp', additions: 2000, deletions: 0 },
+      { filename: 'tests/components/foo/test.esp32-idf.yaml', additions: 5000, deletions: 0 },
+    ]);
+    assert.ok(labels.has('too-big'));
+    assert.equal(labels.size, 1);
+  });
+
+  it('does not label too-big when mega-pr is set', async () => {
+    const labels = await size([
+      { filename: 'esphome/components/foo/foo.cpp', additions: 2000, deletions: 0 },
+    ], true);
+    assert.equal(labels.size, 0);
+  });
+
+  it('produces no size label for a large mega-pr in the gap above medium', async () => {
+    // Non-test changes land between MEDIUM and TOO_BIG: not small/medium, and mega-pr suppresses too-big.
+    const labels = await size([
+      { filename: 'esphome/components/foo/foo.cpp', additions: 500, deletions: 0 },
+    ], true);
+    assert.equal(labels.size, 0);
+  });
+});

.github/scripts/auto-label-pr/tests/reviews.test.js

@@ -0,0 +1,38 @@
+const { describe, it } = require('node:test');
+const assert = require('node:assert/strict');
+const { generateReviewMessages } = require('../reviews');
+
+const MAX_LABELS = 15;
+const TOO_BIG_THRESHOLD = 1000;
+
+function tooBigMessage(labelCount, additions) {
+  const files = [{ filename: 'esphome/components/a/a.cpp', additions, deletions: 0 }];
+  const messages = generateReviewMessages(['too-big'], labelCount, null, files, additions, 0, 'someone', MAX_LABELS, TOO_BIG_THRESHOLD);
+  assert.equal(messages.length, 1);
+  return messages[0];
+}
+
+describe('too-big review message', () => {
+  it('names the line cap when the PR is too long', () => {
+    const message = tooBigMessage(3, 1500);
+    assert.match(message, /will not be reviewed until it is at most 1000 lines \(excluding tests\)\./);
+    assert.match(message, /removed automatically once the PR is at most 1000 lines/);
+    assert.doesNotMatch(message, /components\/areas\./);
+  });
+
+  it('names the area cap when the PR touches too many areas but is short', () => {
+    const message = tooBigMessage(20, 200);
+    assert.match(message, /will not be reviewed until it touches at most 15 components\/areas\./);
+    assert.doesNotMatch(message, /until it is at most 1000 lines/);
+  });
+
+  it('names both caps when both are exceeded', () => {
+    const message = tooBigMessage(20, 1500);
+    assert.match(message, /is at most 1000 lines \(excluding tests\) and touches at most 15 components\/areas\./);
+  });
+
+  it('tells new-component authors to bring one platform first', () => {
+    assert.match(tooBigMessage(3, 1500), /hub and a single platform/);
+    assert.match(tooBigMessage(3, 1500), /mega-pr/);
+  });
+});

.github/scripts/detect-tags.js

@@ -41,7 +41,6 @@ function hasCoreChanges(changedFiles) {
  */
 function hasDashboardChanges(changedFiles) {
   return changedFiles.some(file =>
-    file.startsWith('esphome/dashboard/') ||
     file.startsWith('esphome/components/dashboard_import/')
   );
 }

.github/skills

@@ -0,0 +1 @@
+../.agents/skills

.github/workflows/auto-label-pr.yml

@@ -24,7 +24,7 @@ jobs:
     if: github.event.pull_request.state == 'open' && (github.event.action != 'labeled' || github.event.sender.type != 'Bot')
     steps:
       - name: Checkout
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
 
       - name: Generate a token
         id: generate-token
@@ -35,7 +35,7 @@ jobs:
           # Scope the minted App token to the minimum needed by auto-label-pr/*.js.
           permission-contents: read       # repos.getContent for CODEOWNERS and file lookups in detectors.js
           permission-issues: write        # listLabelsOnIssue, addLabels, removeLabel, list/createComment
-          permission-pull-requests: write  # pulls.listFiles, list/create/update/dismissReview
+          permission-pull-requests: write  # pulls.get, pulls.listFiles, list/create/update/dismissReview
 
       - name: Auto Label PR
         uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0

.github/workflows/ci-api-proto.yml

@@ -21,27 +21,52 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
       - name: Set up Python
-        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
         with:
-          python-version: "3.11"
+          python-version: "3.12"
       - name: Set up uv
         # ``--system`` (below) installs into the setup-python interpreter;
         # no venv is created or restored by this workflow.
-        uses: astral-sh/setup-uv@08807647e7069bb48b6ef5acd8ec9567f424441b # v8.1.0
+        uses: astral-sh/setup-uv@1c37ad07a6a961277cf70c0d37d6f313000f5884 # v10.3.0
         with:
           enable-cache: true
+          # Pull-request-only workflow: a save could never be shared and
+          # would only consume quota.
+          save-cache: "false"
           # Pin uv version so the action does not have to fetch the
           # manifest from raw.githubusercontent.com on every cache
           # miss; that fetch flakes on Windows runners.
           version: "0.11.15"
 
       - name: Install apt dependencies
+        # PR-only workflow, so nothing on dev could seed a shared apt cache
+        # entry; the cached apt action would save one copy per PR. Plain apt
+        # with every call bounded: the apt.conf.d timeouts make a dead
+        # mirror fail over in seconds, and timeout runs under sudo so it can
+        # kill apt-get itself. Install without update first: image lists are
+        # fresh, and the index refresh is what a congested mirror makes slow.
+        timeout-minutes: 15
         run: |
-          sudo apt update
-          sudo apt-cache show protobuf-compiler
-          sudo apt install -y protobuf-compiler
+          sudo tee /etc/apt/apt.conf.d/99ci-acquire-timeouts >/dev/null <<'EOF'
+          Acquire::Retries "1";
+          Acquire::http::Timeout "15";
+          Acquire::https::Timeout "15";
+          EOF
+          # Common path: the image's package lists are fresh enough.
+          if sudo DEBIAN_FRONTEND=noninteractive timeout -k 15 90 \
+              apt-get install -y protobuf-compiler; then
+            protoc --version
+            exit 0
+          fi
+          # Rescue path: refresh the lists once with a generous bound; the
+          # apt config already fails a stalled mirror over quickly.
+          sudo DEBIAN_FRONTEND=noninteractive timeout -k 10 30 \
+            dpkg --configure -a || true
+          sudo timeout -k 15 300 apt-get update
+          sudo DEBIAN_FRONTEND=noninteractive timeout -k 15 300 \
+            apt-get install -y protobuf-compiler
           protoc --version
       - name: Install python dependencies
         run: uv pip install --system aioesphomeapi -c requirements.txt -r requirements_dev.txt
@@ -72,7 +97,7 @@ jobs:
         run: git diff
       - if: failure()
         name: Archive artifacts
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        uses: actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9 # v7.0.2
         with:
           name: generated-proto-files
           path: |

.github/workflows/ci-clang-tidy-hash.yml

@@ -1,76 +0,0 @@
-name: Clang-tidy Hash CI
-
-on:
-  pull_request:
-    paths:
-      - ".clang-tidy"
-      - "platformio.ini"
-      - "requirements_dev.txt"
-      - "sdkconfig.defaults"
-      - ".clang-tidy.hash"
-      - "script/clang_tidy_hash.py"
-      - ".github/workflows/ci-clang-tidy-hash.yml"
-
-permissions:
-  contents: read        # actions/checkout for the PR head
-  pull-requests: write  # pulls.createReview / listReviews / dismissReview when the clang-tidy hash is out of date
-
-jobs:
-  verify-hash:
-    name: Verify clang-tidy hash
-    runs-on: ubuntu-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
-
-      - name: Set up Python
-        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
-        with:
-          python-version: "3.11"
-
-      - name: Verify hash
-        run: |
-          python script/clang_tidy_hash.py --verify
-
-      - if: failure()
-        name: Show hash details
-        run: |
-          python script/clang_tidy_hash.py
-          echo "## Job Failed" | tee -a $GITHUB_STEP_SUMMARY
-          echo "You have modified clang-tidy configuration but have not updated the hash." | tee -a $GITHUB_STEP_SUMMARY
-          echo "Please run 'script/clang_tidy_hash.py --update' and commit the changes." | tee -a $GITHUB_STEP_SUMMARY
-
-      - if: failure() && github.event.pull_request.head.repo.full_name == github.repository
-        name: Request changes
-        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
-        with:
-          script: |
-            await github.rest.pulls.createReview({
-              pull_number: context.issue.number,
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              event: 'REQUEST_CHANGES',
-              body: 'You have modified clang-tidy configuration but have not updated the hash.\nPlease run `script/clang_tidy_hash.py --update` and commit the changes.'
-            })
-
-      - if: success() && github.event.pull_request.head.repo.full_name == github.repository
-        name: Dismiss review
-        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
-        with:
-          script: |
-            let reviews = await github.rest.pulls.listReviews({
-              pull_number: context.issue.number,
-              owner: context.repo.owner,
-              repo: context.repo.repo
-            });
-            for (let review of reviews.data) {
-              if (review.user.login === 'github-actions[bot]' && review.state === 'CHANGES_REQUESTED') {
-                await github.rest.pulls.dismissReview({
-                  pull_number: context.issue.number,
-                  owner: context.repo.owner,
-                  repo: context.repo.repo,
-                  review_id: review.id,
-                  message: 'Clang-tidy hash now matches configuration.'
-                });
-              }
-            }

.github/workflows/ci-docker.yml

@@ -1,28 +1,40 @@
 ---
 name: CI for docker images
 
-# Only run when docker paths change
+# Only run on PRs that touch the docker image, its build inputs, or any code
+# whose toolchain the compile smoke test exercises (core + target platforms).
 
 on:
-  push:
-    branches: [dev, beta, release]
-    paths:
-      - "docker/**"
-      - ".github/workflows/ci-docker.yml"
-      - "requirements*.txt"
-      - "platformio.ini"
-      - "script/platformio_install_deps.py"
-
   pull_request:
     paths:
+      # Docker image and its build inputs.
       - "docker/**"
       - ".github/workflows/ci-docker.yml"
       - "requirements*.txt"
-      - "platformio.ini"
-      - "script/platformio_install_deps.py"
+      - "pyproject.toml"
+      - "esphome/idf_component.yml"
+      # Core, build pipeline, toolchain, and target-platform changes can change
+      # how a toolchain is set up or built, so re-run the per-toolchain compile
+      # smoke test when they change.
+      - "esphome/core/**"
+      - "esphome/writer.py"
+      - "esphome/build_gen/**"
+      - "esphome/build_helpers/**"
+      - "esphome/espidf/**"
+      - "esphome/platformio/**"
+      - "esphome/components/bk72xx/**"
+      - "esphome/components/esp32/**"
+      - "esphome/components/esp8266/**"
+      - "esphome/components/host/**"
+      - "esphome/components/libretiny/**"
+      - "esphome/components/ln882x/**"
+      - "esphome/components/nrf52/**"
+      - "esphome/components/rp2040/**"
+      - "esphome/components/rtl87xx/**"
+      - "esphome/components/zephyr/**"
 
 permissions:
-  contents: read  # actions/checkout only; the build does not push images
+  contents: read  # actions/checkout only
 
 concurrency:
   # yamllint disable-line rule:line-length
@@ -33,6 +45,9 @@ jobs:
   check-docker:
     name: Build docker containers
     runs-on: ${{ matrix.os }}
+    permissions:
+      contents: read   # actions/checkout to load Dockerfile and build context
+      packages: write  # push branch-tagged images to ghcr.io for local testing
     strategy:
       fail-fast: false
       matrix:
@@ -41,23 +56,169 @@ jobs:
           - "ha-addon"
           - "docker"
           # - "lint"
+    outputs:
+      tag: ${{ steps.tag.outputs.tag }}
+      push: ${{ steps.tag.outputs.push }}
     steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
       - name: Set up Python
-        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
         with:
-          python-version: "3.11"
+          python-version: "3.12"
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@d7f5e7f509e45cec5c76c4d5afdd7de93d0b3df5 # v4.1.0
+        uses: docker/setup-buildx-action@594f3bf4285d9ea8dc53c9a0c9c4092420091003 # v4.4.0
 
-      - name: Set TAG
+      - name: Determine tag and whether to push
+        id: tag
+        env:
+          HEAD_REF: ${{ github.head_ref || github.ref_name }}
         run: |
-          echo "TAG=check" >> $GITHUB_ENV
+          # Sanitize the branch name into a valid docker tag: replace invalid
+          # characters, ensure the first character is valid (tags must start
+          # with [A-Za-z0-9_]), and cap the length at 128 characters.
+          branch="$HEAD_REF"
+          tag="${branch//[^a-zA-Z0-9_.-]/-}"
+          case "$tag" in
+            [a-zA-Z0-9_]*) ;;
+            *) tag="pr-${tag}" ;;
+          esac
+          tag="${tag:0:128}"
+          echo "tag=${tag}" >> "$GITHUB_OUTPUT"
+          # Only push branch images for same-repo pull requests. Push events
+          # only fire for dev/beta/release, whose images are owned by the
+          # release pipeline -- never overwrite those from here.
+          if [ "${{ github.event_name }}" = "pull_request" ] \
+            && [ "${{ github.repository }}" = "esphome/esphome" ] \
+            && [ "${{ github.event.pull_request.head.repo.full_name }}" = "esphome/esphome" ]; then
+            echo "push=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "push=false" >> "$GITHUB_OUTPUT"
+          fi
+
+      - name: Log in to the GitHub container registry
+        if: steps.tag.outputs.push == 'true'
+        uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Run build
         run: |
           docker/build.py \
-            --tag "${TAG}" \
+            --tag "${{ steps.tag.outputs.tag }}" \
             --arch "${{ matrix.os == 'ubuntu-24.04-arm' && 'aarch64' || 'amd64' }}" \
             --build-type "${{ matrix.build_type }}" \
-            build
+            --registry ghcr \
+            build ${{ steps.tag.outputs.push == 'true' && '--push --no-cache-to' || '' }} ${{ (matrix.os == 'ubuntu-24.04' && matrix.build_type == 'docker') && '--load' || '' }}
+
+      # The amd64 "docker" image is also loaded locally (above) and handed to
+      # compile-test as an artifact, so the smoke test reuses this build instead
+      # of building the image a second time. Using an artifact (rather than the
+      # pushed image) keeps it working for fork PRs, which never push to ghcr.io.
+      - name: Export image for compile-test
+        if: matrix.os == 'ubuntu-24.04' && matrix.build_type == 'docker'
+        # zstd over gzip: docker save is on the critical path for every
+        # compile-test job, and zstd -T0 is multithreaded (export 50s -> 9s).
+        # docker load auto-detects the format; its time is layer extraction,
+        # not decompression, so it is unchanged. shell: bash adds pipefail so
+        # a failed docker save cannot upload a truncated artifact.
+        shell: bash
+        run: docker save "ghcr.io/esphome/esphome-amd64:${{ steps.tag.outputs.tag }}" | zstd -T0 -3 > compile-test-image.tar.zst
+
+      - name: Upload compile-test image artifact
+        if: matrix.os == 'ubuntu-24.04' && matrix.build_type == 'docker'
+        uses: actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9 # v7.0.2
+        with:
+          # The tar is already compressed, so upload it as-is. archive: false
+          # skips the redundant zip and makes the file name the artifact name
+          # (the `name` input is ignored in that mode).
+          path: compile-test-image.tar.zst
+          retention-days: 1
+          archive: false
+
+  manifest:
+    name: Push ${{ matrix.build_type }} manifest to ghcr.io
+    needs: [check-docker]
+    if: needs.check-docker.outputs.push == 'true'
+    runs-on: ubuntu-24.04
+    permissions:
+      contents: read   # actions/checkout to run docker/build.py
+      packages: write  # buildx imagetools writes the multi-arch tag to ghcr.io
+    strategy:
+      fail-fast: false
+      matrix:
+        build_type:
+          - "ha-addon"
+          - "docker"
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - name: Set up Python
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
+        with:
+          python-version: "3.12"
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@594f3bf4285d9ea8dc53c9a0c9c4092420091003 # v4.4.0
+
+      - name: Log in to the GitHub container registry
+        uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Create and push manifest
+        run: |
+          docker/build.py \
+            --tag "${{ needs.check-docker.outputs.tag }}" \
+            --build-type "${{ matrix.build_type }}" \
+            --registry ghcr \
+            manifest
+
+  # Smoke-test the built image by compiling one minimal config per target
+  # platform / toolchain. This catches missing system dependencies in the image
+  # that only surface when a given toolchain is downloaded and run. The image is
+  # the amd64 "docker" build produced by check-docker (shared as an artifact).
+  compile-test:
+    name: Compile ${{ matrix.id }}
+    needs: check-docker
+    runs-on: ubuntu-24.04
+    permissions:
+      contents: read  # actions/checkout to load the test configs
+    strategy:
+      fail-fast: false
+      matrix:
+        # One entry per distinct toolchain. ESP32 variants (c3/c6/s2/s3/p4)
+        # share a toolchain bundle, so esp32 is exercised on the base variant
+        # across the full framework x toolchain cross-product (arduino/esp-idf
+        # framework, each built with the platformio and native esp-idf
+        # toolchains) so both toolchains stay covered regardless of which one is
+        # the default.
+        id:
+          - esp8266-arduino
+          - esp8266-arduino-native
+          - esp32-arduino-platformio
+          - esp32-arduino-esp-idf
+          - esp32-idf-platformio
+          - esp32-idf-esp-idf
+          - rp2040-arduino
+          - bk72xx-arduino
+          - rtl87xx-arduino
+          - ln882x-arduino
+          - nrf52
+          - host
+    steps:
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+      - name: Download image artifact
+        uses: actions/download-artifact@9000827ccba6bdab643e8b6fd33ac0654aef8333 # v8.0.2
+        with:
+          name: compile-test-image.tar.zst
+      - name: Load image
+        run: docker load --input compile-test-image.tar.zst
+      - name: Compile ${{ matrix.id }}
+        run: |
+          docker run --rm \
+            -v "${{ github.workspace }}/docker/test_configs:/config" \
+            -e ESPHOME_LDGEN_STRICT=1 \
+            "ghcr.io/esphome/esphome-amd64:${{ needs.check-docker.outputs.tag }}" \
+            compile "${{ matrix.id }}.yaml"

.github/workflows/ci-github-scripts.yml

@@ -20,7 +20,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Check out code from GitHub
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
 
       - name: Run tests
         working-directory: .github/scripts/auto-label-pr

.github/workflows/ci-memory-impact-comment.yml

@@ -49,7 +49,7 @@ jobs:
 
       - name: Check out code from base repository
         if: steps.pr.outputs.skip != 'true'
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           # Always check out from the base repository (esphome/esphome), never from forks
           # Use the PR's target branch to ensure we run trusted code from the main repo
@@ -60,7 +60,7 @@ jobs:
         if: steps.pr.outputs.skip != 'true'
         uses: ./.github/actions/restore-python
         with:
-          python-version: "3.11"
+          python-version: "3.12"
           cache-key: ${{ hashFiles('.cache-key') }}
 
       - name: Download memory analysis artifacts

.github/workflows/codeowner-approved-label-update.yml

@@ -26,7 +26,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout base branch
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           ref: ${{ github.event.pull_request.base.sha }}
           sparse-checkout: |

.github/workflows/codeowner-review-request.yml

@@ -29,7 +29,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout base branch
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           ref: ${{ github.event.pull_request.base.sha }}
 

.github/workflows/codeql.yml

@@ -52,11 +52,11 @@ jobs:
             # your codebase is analyzed, see https://docs.github.com/en/code-security/code-scanning/creating-an-advanced-setup-for-code-scanning/codeql-code-scanning-for-compiled-languages
     steps:
       - name: Checkout repository
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
 
       # Initializes the CodeQL tools for scanning.
       - name: Initialize CodeQL
-        uses: github/codeql-action/init@7211b7c8077ea37d8641b6271f6a365a22a5fbfa # v4.36.0
+        uses: github/codeql-action/init@24c54180a607b1449ed407dd24f251e4e9147c8d # v4.38.3
         with:
           languages: ${{ matrix.language }}
           build-mode: ${{ matrix.build-mode }}
@@ -84,6 +84,6 @@ jobs:
           exit 1
 
       - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@7211b7c8077ea37d8641b6271f6a365a22a5fbfa # v4.36.0
+        uses: github/codeql-action/analyze@24c54180a607b1449ed407dd24f251e4e9147c8d # v4.38.3
         with:
           category: "/language:${{matrix.language}}"

.github/workflows/dashboard-deprecation-comment.yml

@@ -1,119 +0,0 @@
-name: Add Dashboard Deprecation Comment
-
-on:
-  pull_request_target:
-    types: [opened, synchronize]
-
-# All API calls (pulls.listFiles + issues.{list,create,update}Comment) are performed with
-# the App token minted below, so the workflow's GITHUB_TOKEN does not need any scopes.
-permissions: {}
-
-jobs:
-  dashboard-deprecation-comment:
-    name: Dashboard deprecation comment
-    runs-on: ubuntu-latest
-    # Release-bump PRs (bump-X.Y.Z -> beta, beta -> release) inevitably
-    # roll up everything merged into dev since the last cut, which can
-    # include dashboard changes that have already been reviewed once.
-    # The bot's purpose is to warn new contributors before they invest
-    # time -- that only applies to PRs entering dev.
-    if: github.event.pull_request.base.ref == 'dev'
-    steps:
-      - name: Generate a token
-        id: generate-token
-        uses: actions/create-github-app-token@bcd2ba49218906704ab6c1aa796996da409d3eb1 # v3.2.0
-        with:
-          client-id: ${{ vars.ESPHOME_GITHUB_APP_CLIENT_ID }}
-          private-key: ${{ secrets.ESPHOME_GITHUB_APP_PRIVATE_KEY }}
-          # pulls.listFiles + issues.{list,create,update}Comment on PRs. For PR resources
-          # the issues.*Comment APIs require the pull-requests scope, not issues.
-          permission-pull-requests: write
-
-      - name: Add dashboard deprecation comment
-        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
-        with:
-          github-token: ${{ steps.generate-token.outputs.token }}
-          script: |
-            const commentMarker = "<!-- This comment was generated automatically by the dashboard-deprecation-comment workflow. -->";
-
-            const commentBody = `Thanks for opening this PR!
-
-            Heads up: the legacy ESPHome dashboard (\`esphome/dashboard/\` and \`tests/dashboard/\`) is **deprecated** and is being replaced by [ESPHome Device Builder](https://github.com/esphome/device-builder). We are not adding new features to the legacy dashboard and it will eventually be removed from this repository.
-
-            What this means for your PR:
-
-            - **New features / enhancements**: please port the change to [esphome/device-builder](https://github.com/esphome/device-builder) instead. We are unlikely to review or merge new dashboard features here.
-            - **Bug fixes**: small fixes may still be considered, but please check first whether the same issue exists in Device Builder, where the fix will have a longer life.
-            - **Security issues**: please do not file a public PR. Report privately via [GitHub security advisories](https://github.com/esphome/esphome/security/advisories/new) so we can coordinate a fix.
-
-            We appreciate the contribution and apologize for the friction; flagging this early so your time isn't spent on a change that may not land.
-
-            ---
-            (Added by the PR bot)
-
-            ${commentMarker}`;
-
-            async function getDashboardChanges(github, owner, repo, prNumber) {
-                const changedFiles = await github.paginate(
-                    github.rest.pulls.listFiles,
-                    {
-                        owner: owner,
-                        repo: repo,
-                        pull_number: prNumber,
-                        per_page: 100,
-                    }
-                );
-
-                return changedFiles.filter(file =>
-                    file.filename.startsWith('esphome/dashboard/') ||
-                    file.filename.startsWith('tests/dashboard/')
-                );
-            }
-
-            async function findBotComment(github, owner, repo, prNumber) {
-                const comments = await github.paginate(
-                    github.rest.issues.listComments,
-                    {
-                        owner: owner,
-                        repo: repo,
-                        issue_number: prNumber,
-                        per_page: 100,
-                    }
-                );
-
-                return comments.find(comment =>
-                    comment.body.includes(commentMarker) && comment.user.type === "Bot"
-                );
-            }
-
-            const prNumber = context.payload.pull_request.number;
-            const { owner, repo } = context.repo;
-
-            const dashboardChanges = await getDashboardChanges(github, owner, repo, prNumber);
-            const existingComment = await findBotComment(github, owner, repo, prNumber);
-
-            if (dashboardChanges.length === 0) {
-                // PR doesn't (or no longer) touches the legacy dashboard. If we previously
-                // commented (e.g. files were removed in a later push), leave the comment in
-                // place for history rather than thrash on edit/delete.
-                return;
-            }
-
-            if (existingComment) {
-                if (existingComment.body === commentBody) {
-                    return;
-                }
-                await github.rest.issues.updateComment({
-                    owner: owner,
-                    repo: repo,
-                    comment_id: existingComment.id,
-                    body: commentBody,
-                });
-            } else {
-                await github.rest.issues.createComment({
-                    owner: owner,
-                    repo: repo,
-                    issue_number: prNumber,
-                    body: commentBody,
-                });
-            }

.github/workflows/external-component-bot.yml

@@ -125,14 +125,20 @@ jobs:
             }
 
             async function getEsphomeAndComponentChanges(github, owner, repo, prNumber) {
-                const changedFiles = await github.rest.pulls.listFiles({
-                    owner: owner,
-                    repo: repo,
-                    pull_number: prNumber,
-                });
+                const changedFiles = await github.paginate(
+                    github.rest.pulls.listFiles,
+                    {
+                        owner: owner,
+                        repo: repo,
+                        pull_number: prNumber,
+                        per_page: 100,
+                    }
+                );
 
-                const esphomeChanges = changedFiles.data
-                    .filter(file => file.filename !== "esphome/core/defines.h" && file.filename.startsWith('esphome/'))
+                // Files used only for development and CI, which do not affect use as an external component
+                const ignoredFiles = ["esphome/core/defines.h", "esphome/idf_component.yml"];
+                const esphomeChanges = changedFiles
+                    .filter(file => !ignoredFiles.includes(file.filename) && file.filename.startsWith('esphome/'))
                     .map(file => {
                         const match = file.filename.match(/esphome\/([^/]+)/);
                         return match ? match[1] : null;
@@ -144,7 +150,7 @@ jobs:
                 }
 
                 const uniqueEsphomeChanges = [...new Set(esphomeChanges)];
-                const componentChanges = changedFiles.data
+                const componentChanges = changedFiles
                     .filter(file => file.filename.startsWith('esphome/components/'))
                     .map(file => {
                         const match = file.filename.match(/esphome\/components\/([^/]+)\//);

.github/workflows/lock.yml

@@ -14,4 +14,4 @@ jobs:
     permissions:
       issues: write         # issues.lock on closed issues
       pull-requests: write  # issues.lock on closed pull requests
-    uses: esphome/workflows/.github/workflows/lock.yml@025a1e6255610c498ed590403b7e510b69e474df  # 2026.4.1
+    uses: esphome/workflows/.github/workflows/lock.yml@cc3e76de337dc59bc1cba8da58d963cd23b873f1  # 2026.9.0

.github/workflows/pr-title-check.yml

@@ -16,7 +16,7 @@ jobs:
     name: Validate PR title
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
 
       - uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
         with:

.github/workflows/release-nightly.yml

@@ -0,0 +1,38 @@
+---
+name: Nightly Dev Release
+
+# Works out the dated dev tag and starts the release workflow with it, so that
+# the release run is named after the tag it builds. A workflow run name is
+# fixed when the run starts and cannot read a file or the current date.
+
+on:
+  schedule:
+    - cron: "0 2 * * *"
+
+permissions:
+  contents: read  # actions/checkout to read the version from esphome/const.py
+
+jobs:
+  trigger:
+    name: Start release build
+    if: github.repository == 'esphome/esphome'
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read  # actions/checkout to read the version from esphome/const.py
+      actions: write  # gh workflow run starts release.yml
+    steps:
+      - name: Check out the repository
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+
+      - name: Start the release workflow
+        env:
+          GH_TOKEN: ${{ github.token }}
+        run: |
+          VERSION=$(sed -n -E "s/^__version__\s+=\s+\"(.+)\"$/\1/p" esphome/const.py)
+          if [[ -z "$VERSION" ]]; then
+            echo "::error::Could not read __version__ from esphome/const.py"
+            exit 1
+          fi
+          TAG="${VERSION}$(date --utc '+%Y%m%d')"
+          echo "Starting release build for ${TAG}"
+          gh workflow run release.yml --ref "${GITHUB_REF_NAME}" --field tag="${TAG}"

.github/workflows/release.yml

@@ -1,12 +1,23 @@
 ---
 name: Publish Release
 
+# Releases (production and beta) are named after the version they publish.
+# Dev builds are named after the dated dev tag, which is passed in by the
+# nightly workflow because a run name cannot compute it itself.
+run-name: ${{ github.event.inputs.tag || github.event.release.tag_name || format('Manual build ({0})', github.ref_name) }}
+
 on:
   workflow_dispatch:
+    inputs:
+      tag:
+        description: >-
+          Tag to build. Only supported on dev, where the nightly workflow
+          uses it. Leave empty to build the version from esphome/const.py
+          with today's date appended.
+        required: false
+        default: ""
   release:
     types: [published]
-  schedule:
-    - cron: "0 2 * * *"
 
 permissions:
   contents: read  # actions/checkout for all jobs; deploy jobs add their own scopes when they need to write
@@ -20,9 +31,11 @@ jobs:
       branch_build: ${{ steps.tag.outputs.branch_build }}
       deploy_env: ${{ steps.tag.outputs.deploy_env }}
     steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
       - name: Get tag
         id: tag
+        env:
+          INPUT_TAG: ${{ github.event.inputs.tag }}
         # yamllint disable rule:line-length
         run: |
           if [[ "${{ github.event_name }}" = "release" ]]; then
@@ -34,12 +47,23 @@ jobs:
               ENVIRONMENT="production"
             fi
           else
-            TAG=$(cat esphome/const.py | sed -n -E "s/^__version__\s+=\s+\"(.+)\"$/\1/p")
-            today="$(date --utc '+%Y%m%d')"
-            TAG="${TAG}${today}"
             BRANCH=${GITHUB_REF#refs/heads/}
+            # The nightly workflow passes the finished tag so that the run name
+            # matches what is built. Without it, work it out here.
+            TAG="${INPUT_TAG}"
+            if [[ -n "$TAG" && "$BRANCH" != "dev" ]]; then
+              echo "::error::The tag input is only supported on dev. A build from ${BRANCH} has to use the tag worked out here, which carries the branch name, so that it cannot publish over the dev, beta, latest or stable images."
+              exit 1
+            fi
+            if [[ -z "$TAG" ]]; then
+              TAG=$(cat esphome/const.py | sed -n -E "s/^__version__\s+=\s+\"(.+)\"$/\1/p")
+              today="$(date --utc '+%Y%m%d')"
+              TAG="${TAG}${today}"
+              if [[ "$BRANCH" != "dev" ]]; then
+                TAG="${TAG}-${BRANCH}"
+              fi
+            fi
             if [[ "$BRANCH" != "dev" ]]; then
-              TAG="${TAG}-${BRANCH}"
               BRANCH_BUILD="true"
               ENVIRONMENT=""
             else
@@ -60,9 +84,9 @@ jobs:
       contents: read   # actions/checkout to build the sdist/wheel
       id-token: write  # OIDC token for PyPI Trusted Publishing (pypa/gh-action-pypi-publish)
     steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
       - name: Set up Python
-        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
         with:
           python-version: "3.x"
       - name: Build
@@ -70,7 +94,7 @@ jobs:
           pip3 install build
           python3 -m build
       - name: Publish
-        uses: pypa/gh-action-pypi-publish@cef221092ed1bacb1cc03d23a2d87d1d172e277b # v1.14.0
+        uses: pypa/gh-action-pypi-publish@dc37677b2e1c63e2034f94d8a5b11f265b73ba33 # v1.14.2
         with:
           skip-existing: true
 
@@ -92,22 +116,22 @@ jobs:
             os: "ubuntu-24.04-arm"
 
     steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
       - name: Set up Python
-        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
         with:
-          python-version: "3.11"
+          python-version: "3.12"
 
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@d7f5e7f509e45cec5c76c4d5afdd7de93d0b3df5 # v4.1.0
+        uses: docker/setup-buildx-action@594f3bf4285d9ea8dc53c9a0c9c4092420091003 # v4.4.0
 
       - name: Log in to docker hub
-        uses: docker/login-action@650006c6eb7dba73a995cc03b0b2d7f5ca915bee # v4.2.0
+        uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
         with:
           username: ${{ secrets.DOCKER_USER }}
           password: ${{ secrets.DOCKER_PASSWORD }}
       - name: Log in to the GitHub container registry
-        uses: docker/login-action@650006c6eb7dba73a995cc03b0b2d7f5ca915bee # v4.2.0
+        uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
@@ -138,7 +162,7 @@ jobs:
       #     version: ${{ needs.init.outputs.tag }}
 
       - name: Upload digests
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
+        uses: actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9 # v7.0.2
         with:
           name: digests-${{ matrix.platform.arch }}
           path: /tmp/digests
@@ -168,27 +192,27 @@ jobs:
           - ghcr
           - dockerhub
     steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+      - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
 
       - name: Download digests
-        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
+        uses: actions/download-artifact@9000827ccba6bdab643e8b6fd33ac0654aef8333 # v8.0.2
         with:
           pattern: digests-*
           path: /tmp/digests
           merge-multiple: true
 
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@d7f5e7f509e45cec5c76c4d5afdd7de93d0b3df5 # v4.1.0
+        uses: docker/setup-buildx-action@594f3bf4285d9ea8dc53c9a0c9c4092420091003 # v4.4.0
 
       - name: Log in to docker hub
         if: matrix.registry == 'dockerhub'
-        uses: docker/login-action@650006c6eb7dba73a995cc03b0b2d7f5ca915bee # v4.2.0
+        uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
         with:
           username: ${{ secrets.DOCKER_USER }}
           password: ${{ secrets.DOCKER_PASSWORD }}
       - name: Log in to the GitHub container registry
         if: matrix.registry == 'ghcr'
-        uses: docker/login-action@650006c6eb7dba73a995cc03b0b2d7f5ca915bee # v4.2.0
+        uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
         with:
           registry: ghcr.io
           username: ${{ github.actor }}

.github/workflows/stale.yml

@@ -6,61 +6,60 @@ on:
     - cron: "30 0 * * *"
   workflow_dispatch:
 
-permissions:
-  issues: write         # actions/stale labels, comments on, and closes stale issues
-  pull-requests: write  # actions/stale labels, comments on, and closes stale pull requests
-
-concurrency:
-  group: lock
+# The reusable workflow authenticates as the ESPHome GitHub App, so GITHUB_TOKEN
+# needs no permissions at all.
+permissions: {}
 
 jobs:
   stale:
     if: github.repository_owner == 'esphome'
-    runs-on: ubuntu-latest
-    steps:
-      - name: Stale
-        uses: actions/stale@eb5cf3af3ac0a1aa4c9c45633dd1ae542a27a899 # v10.3.0
-        with:
-          debug-only: ${{ github.ref != 'refs/heads/dev' }} # Dry-run when not run on dev branch
-          remove-stale-when-updated: true
-          operations-per-run: 400
+    # No GITHUB_TOKEN permissions: the reusable workflow mints an ESPHome
+    # GitHub App token so the labels, comments and closures come from
+    # esphome[bot] instead of github-actions[bot].
+    uses: esphome/workflows/.github/workflows/stale.yml@4fbce13026c393d6eb6b247e394dcd9ef1af76c2  # main
+    secrets:
+      ESPHOME_GITHUB_APP_PRIVATE_KEY: ${{ secrets.ESPHOME_GITHUB_APP_PRIVATE_KEY }}
+    with:
+      # Live only on dev: a workflow_dispatch from any other branch is a dry run
+      dry-run: ${{ github.ref != 'refs/heads/dev' }}
+      days-before-stale: 90
+      days-before-stale-conflicting: 30
+      days-before-close: 7
+      stale-label: stale
+      exempt-label: not-stale
+      conflict-label: needs-rebase
+      ignored-users: esphbot,codecov-commenter
+      stale-pr-message: >
+        There hasn't been any activity on this pull request recently. This
+        pull request has been automatically marked as stale because of that
+        and will be closed if no further activity occurs within 7 days.
 
-          # The 90 day stale policy for PRs
-          # - PRs
-          # - No PRs marked as "not-stale"
-          # - No Issues (see below)
-          days-before-pr-stale: 90
-          days-before-pr-close: 7
-          stale-pr-label: "stale"
-          exempt-pr-labels: "not-stale"
-          stale-pr-message: >
-            There hasn't been any activity on this pull request recently. This
-            pull request has been automatically marked as stale because of that
-            and will be closed if no further activity occurs within 7 days.
+        If you are the author of this PR, please leave a comment if you want
+        to keep it open. Also, please merge the latest dev branch into your
+        branch to ensure that it's up to date with the latest changes.
 
-            If you are the author of this PR, please leave a comment if you want
-            to keep it open. Also, please rebase your PR onto the latest dev
-            branch to ensure that it's up to date with the latest changes.
+        Thank you for your contribution!
+      conflict-pr-message: >
+        This pull request has merge conflicts. Please merge the latest dev
+        branch into your branch to resolve them.
+      stale-conflicting-pr-message: >
+        The merge conflicts on this pull request have not been resolved for
+        30 days, so it has been marked as stale and will be closed if they
+        are not resolved within 7 days.
 
-            Thank you for your contribution!
+        Please merge the latest dev branch into your branch. Once the pull
+        request is mergeable again the stale label is removed.
 
-          # The 90 day stale policy for Issues
-          # - Issues
-          # - No Issues marked as "not-stale"
-          # - No PRs (see above)
-          days-before-issue-stale: 90
-          days-before-issue-close: 7
-          stale-issue-label: "stale"
-          exempt-issue-labels: "not-stale"
-          stale-issue-message: >
-            There hasn't been any activity on this issue recently. Due to the
-            high number of incoming GitHub notifications, we have to clean some
-            of the old issues, as many of them have already been resolved with
-            the latest updates.
+        Thank you for your contribution!
+      stale-issue-message: >
+        There hasn't been any activity on this issue recently. Due to the
+        high number of incoming GitHub notifications, we have to clean some
+        of the old issues, as many of them have already been resolved with
+        the latest updates.
 
-            Please make sure to update to the latest ESPHome version and
-            check if that solves the issue. Let us know if that works for you by
-            adding a comment 👍
+        Please make sure to update to the latest ESPHome version and
+        check if that solves the issue. Let us know if that works for you by
+        adding a comment 👍
 
-            This issue has now been marked as stale and will be closed if no
-            further activity occurs. Thank you for your contributions.
+        This issue has now been marked as stale and will be closed if no
+        further activity occurs. Thank you for your contributions.

.github/workflows/status-check-labels.yml

@@ -5,11 +5,7 @@ on:
     types: [opened, reopened, labeled, unlabeled, synchronize]
 
 permissions:
-  pull-requests: read  # issues.listLabelsOnIssue to detect blocking labels (needs-docs, merge-after-release, chained-pr)
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.event.pull_request.number }}
-  cancel-in-progress: true
+  pull-requests: read  # issues.listLabelsOnIssue to detect blocking labels (needs-docs, needs-developer-docs, merge-after-release, chained-pr)
 
 jobs:
   check:
@@ -20,7 +16,7 @@ jobs:
         uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
         with:
           script: |
-            const blockingLabels = ['needs-docs', 'merge-after-release', 'chained-pr'];
+            const blockingLabels = ['needs-docs', 'needs-developer-docs', 'merge-after-release', 'chained-pr'];
             const { data: labels } = await github.rest.issues.listLabelsOnIssue({
               owner: context.repo.owner,
               repo: context.repo.repo,

.github/workflows/sync-dependency-versions.yml

@@ -0,0 +1,96 @@
+# Keeps pre-commit hook revs in sync with the requirements files.
+#
+# Dependabot only bumps the pins in requirements*.txt. Some of those tools
+# are pinned again as hook revs in .pre-commit-config.yaml. This workflow
+# runs script/sync_dependency_versions.py against the pull request branch
+# and pushes a commit with the revs updated.
+
+name: Sync dependency versions
+
+on:
+  # pull_request_target rather than pull_request so the App secret is
+  # available on Dependabot pull requests (pull_request runs opened by
+  # Dependabot only see Dependabot secrets). The job below only touches
+  # branches in this repository and only ever executes the script from the
+  # base branch checkout, so fork code never runs with the token.
+  pull_request_target:
+    types: [opened, synchronize, reopened]
+    branches:
+      - dev
+    paths:
+      - requirements_dev.txt
+      - requirements_test.txt
+      - .pre-commit-config.yaml
+      - script/sync_dependency_versions.py
+
+# The push to the pull request branch uses the App token minted below, so
+# the workflow's GITHUB_TOKEN does not need any scopes.
+permissions: {}
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number }}
+  cancel-in-progress: true
+
+jobs:
+  sync:
+    name: Sync pinned versions
+    runs-on: ubuntu-latest
+    # Same-repository branches only: a push to a fork is not possible with
+    # this token, and it keeps untrusted heads out of a privileged job.
+    if: >-
+      github.repository == 'esphome/esphome'
+      && github.event.pull_request.head.repo.full_name == github.repository
+    steps:
+      - name: Generate a token
+        id: generate-token
+        uses: actions/create-github-app-token@bcd2ba49218906704ab6c1aa796996da409d3eb1 # v3.2.0
+        with:
+          client-id: ${{ vars.ESPHOME_GITHUB_APP_CLIENT_ID }}
+          private-key: ${{ secrets.ESPHOME_GITHUB_APP_PRIVATE_KEY }}
+          # A push made with the workflow's own GITHUB_TOKEN would not start
+          # CI on the new commit; a push with the App token does.
+          permission-contents: write  # git push of the sync commit to the pull request branch
+
+      - name: Check out base branch
+        # Provides the script that runs below. Deliberately the base branch
+        # so the pull request cannot change what executes here.
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          ref: ${{ github.event.pull_request.base.sha }}
+          persist-credentials: false
+
+      - name: Check out pull request branch
+        # No allow-unsafe-pr-checkout here on purpose: checkout v7 only
+        # refuses heads that live in a different repository, and the job
+        # condition above already limits runs to same-repository branches.
+        # Leaving it off keeps that refusal as a backstop for fork heads.
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+        with:
+          ref: ${{ github.event.pull_request.head.ref }}
+          path: pull-request
+          token: ${{ steps.generate-token.outputs.token }}
+
+      - name: Set up Python
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
+        with:
+          python-version: "3.12"
+
+      - name: Install yamlrocks
+        # The script edits YAML through yamlrocks. Take the pin from the
+        # base branch requirements so this workflow has no copy of its own.
+        run: pip install "$(grep -E '^yamlrocks==' requirements_test.txt | cut -d'#' -f1)"
+
+      - name: Sync pinned versions
+        run: python script/sync_dependency_versions.py --root pull-request
+
+      - name: Push changes
+        working-directory: pull-request
+        run: |
+          if git diff --quiet; then
+            echo "All pinned versions already match the requirements files."
+            exit 0
+          fi
+          git config user.name "esphome[bot]"
+          git config user.email "115708604+esphome[bot]@users.noreply.github.com"
+          git commit -am "Sync pinned tool versions with requirements files"
+          git push

.github/workflows/sync-device-classes.yml

@@ -28,26 +28,26 @@ jobs:
           permission-pull-requests: write  # pulls.create / pulls.update to open or refresh the sync PR
 
       - name: Checkout
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
 
       - name: Checkout Home Assistant
-        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
           repository: home-assistant/core
           path: lib/home-assistant
 
       - name: Setup Python
-        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
         with:
           python-version: "3.14"
 
       - name: Set up uv
         # An order of magnitude faster than pip on cold boots, with its
         # own wheel cache. ``--system`` (below) installs into the
-        # setup-python interpreter so subsequent ``pre-commit`` /
+        # setup-python interpreter so subsequent ``prek`` /
         # ``script/run-in-env.py`` steps find the deps without a
         # ``uv run`` prefix.
-        uses: astral-sh/setup-uv@08807647e7069bb48b6ef5acd8ec9567f424441b # v8.1.0
+        uses: astral-sh/setup-uv@1c37ad07a6a961277cf70c0d37d6f313000f5884 # v10.3.0
         with:
           enable-cache: true
           # Pin uv version so the action does not have to fetch the
@@ -58,19 +58,19 @@ jobs:
       - name: Install Home Assistant
         run: |
           uv pip install --system -e lib/home-assistant
-          uv pip install --system -r requirements.txt -r requirements_test.txt pre-commit
+          uv pip install --system -r requirements.txt -r requirements_test.txt
 
       - name: Sync
         run: |
           python ./script/sync-device_class.py
 
-      - name: Apply pre-commit auto-fixes
+      - name: Apply prek auto-fixes
         # First pass: let formatters (ruff, end-of-file-fixer, etc.) modify
-        # files. pre-commit exits non-zero whenever a hook touches anything,
+        # files. prek exits non-zero whenever a hook touches anything,
         # which would otherwise abort the workflow before the auto-fixes
         # can flow into the sync PR.
         #
-        # SKIP:
+        # PREK_SKIP:
         #   - no-commit-to-branch is a local guard against committing on
         #     dev/release/beta; CI runs on dev by definition, and
         #     peter-evans/create-pull-request creates the branch itself.
@@ -79,25 +79,25 @@ jobs:
         #     the runtime deps (HA + requirements*.txt); main CI already
         #     gates pylint on real PRs.
         env:
-          SKIP: pylint,no-commit-to-branch
-        run: python script/run-in-env.py pre-commit run --all-files || true
+          PREK_SKIP: pylint,no-commit-to-branch
+        run: python script/run-in-env.py prek run --all-files || true
 
-      - name: Verify pre-commit clean
+      - name: Verify prek clean
         # Second pass: re-run all hooks against the now-fixed tree.
         # Auto-fixers exit 0 (nothing to change); any remaining failure
         # from a check-only hook (flake8 / yamllint / ci-custom) is a
-        # real issue and fails the workflow loudly. Same SKIP list as
+        # real issue and fails the workflow loudly. Same PREK_SKIP list as
         # above for the same reasons.
         env:
-          SKIP: pylint,no-commit-to-branch
-        run: python script/run-in-env.py pre-commit run --all-files
+          PREK_SKIP: pylint,no-commit-to-branch
+        run: python script/run-in-env.py prek run --all-files
 
       - name: Commit changes
         uses: peter-evans/create-pull-request@5f6978faf089d4d20b00c7766989d076bb2fc7f1 # v8.1.1
         with:
           commit-message: "Synchronise Device Classes from Home Assistant"
-          committer: esphomebot <esphome@openhomefoundation.org>
-          author: esphomebot <esphome@openhomefoundation.org>
+          committer: esphome[bot] <115708604+esphome[bot]@users.noreply.github.com>
+          author: esphome[bot] <115708604+esphome[bot]@users.noreply.github.com>
           branch: sync/device-classes
           delete-branch: true
           title: "Synchronise Device Classes from Home Assistant"

.github/workflows/sync-integration-durations.yml

@@ -0,0 +1,98 @@
+---
+name: Refresh integration test durations
+
+on:
+  workflow_dispatch:
+  schedule:
+    - cron: "45 5 * * 1"
+
+# Repo writes (branch push, PR open) happen via the App token minted below,
+# so the workflow's GITHUB_TOKEN does not need any write scopes.
+permissions:
+  contents: read
+  actions: read  # gh api / gh run download for the CI junit artifacts
+
+jobs:
+  sync:
+    name: Refresh integration test durations
+    runs-on: ubuntu-latest
+    if: github.repository == 'esphome/esphome'
+    steps:
+      - name: Generate a token
+        id: generate-token
+        uses: actions/create-github-app-token@bcd2ba49218906704ab6c1aa796996da409d3eb1 # v3.2.0
+        with:
+          client-id: ${{ vars.ESPHOME_GITHUB_APP_CLIENT_ID }}
+          private-key: ${{ secrets.ESPHOME_GITHUB_APP_PRIVATE_KEY }}
+          permission-contents: write       # push the sync branch
+          permission-pull-requests: write  # open or refresh the sync PR
+
+      - name: Checkout
+        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
+
+      - name: Setup Python
+        uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
+        with:
+          python-version: "3.13"
+
+      - name: Refresh from the newest usable dev run
+        env:
+          GH_TOKEN: ${{ github.token }}
+        run: |
+          # Only full matrix dev runs upload junit-integration-* artifacts
+          # (see the integration-tests job); the merge script re-checks
+          # coverage regardless.
+          # Newest-first candidates via their bucket-0 artifact. Fork PRs run
+          # their own ci.yml, so name and branch are spoofable; require
+          # same-repo. Assignment failures trip set -e and fail loudly.
+          candidates=$(
+            gh api "repos/${GITHUB_REPOSITORY}/actions/artifacts?name=junit-integration-0&per_page=100" \
+              --jq '.artifacts[] | select(.expired | not) | .workflow_run
+                    | select(.head_branch == "dev" and .head_repository_id != null
+                             and .head_repository_id == .repository_id)
+                    | .id'
+          )
+          # Green runs first, then the rest newest first; a run missing a
+          # bucket fails the coverage check and the next one is tried
+          green=""
+          rest=""
+          for id in ${candidates}; do
+            conclusion=$(gh api "repos/${GITHUB_REPOSITORY}/actions/runs/${id}" --jq '.conclusion // ""')
+            if [ "${conclusion}" = "success" ]; then
+              green="${green} ${id}"
+            elif [ -n "${conclusion}" ]; then
+              rest="${rest} ${id}"
+            fi
+          done
+          # helpers.py imports colorama; the script needs nothing else
+          pip install colorama
+          for id in ${green} ${rest}; do
+            rm -rf /tmp/junit
+            if ! gh run download "${id}" --repo "${GITHUB_REPOSITORY}" -p "junit-integration-*" -D /tmp/junit; then
+              echo "::warning::Could not download artifacts for run ${id}; trying the next"
+              continue
+            fi
+            status=0
+            python script/update_integration_test_durations.py /tmp/junit || status=$?
+            if [ "${status}" -eq 0 ]; then
+              echo "Refreshed from run ${id}"
+              exit 0
+            fi
+            # Only EXIT_LOW_COVERAGE (3) from the script advances to the next run
+            [ "${status}" -eq 3 ] || exit 1
+            echo "::warning::Run ${id} covers too few test files; trying the next"
+          done
+          echo "::error::No dev CI run with usable junit artifacts in range; the feed is starved"
+          exit 1
+
+      - name: Commit changes
+        uses: peter-evans/create-pull-request@5f6978faf089d4d20b00c7766989d076bb2fc7f1 # v8.1.1
+        with:
+          commit-message: "[ci] Refresh integration test durations"
+          committer: esphome[bot] <115708604+esphome[bot]@users.noreply.github.com>
+          author: esphome[bot] <115708604+esphome[bot]@users.noreply.github.com>
+          branch: sync/integration-durations
+          delete-branch: true
+          title: "[ci] Refresh integration test durations"
+          body-path: .github/PULL_REQUEST_TEMPLATE.md
+          token: ${{ steps.generate-token.outputs.token }}

.gitignore

@@ -133,14 +133,19 @@ CTestTestfile.cmake
 .gcc-flags.json
 
 config/
+# Test fixture config/ directories are tracked (the rule above is the dashboard dir)
+!tests/component_tests/**/config/
 tests/build/
 tests/.esphome/
+# Output kept by failing tests for inspection; uploaded by CI
+test_artifacts/
 /.temp-clang-tidy.cpp
 /.temp/
 .pio/
 
 sdkconfig.*
 !sdkconfig.defaults
+!sdkconfig.defaults.*
 
 .tests/
 

.pre-commit-config.yaml

@@ -1,17 +1,16 @@
 ---
 # See https://pre-commit.com for more information
 # See https://pre-commit.com/hooks.html for more hooks
-
 ci:
   autoupdate_commit_msg: 'pre-commit: autoupdate'
   autoupdate_schedule: off  # Disabled until ruff versions are synced between deps and pre-commit
   # Skip hooks that have issues in pre-commit CI environment
-  skip: [pylint, clang-tidy-hash]
+  skip: [pylint]
 
 repos:
   - repo: https://github.com/astral-sh/ruff-pre-commit
     # Ruff version.
-    rev: v0.15.15
+    rev: v0.17.0
     hooks:
       # Run the linter.
       - id: ruff
@@ -19,7 +18,7 @@ repos:
       # Run the formatter.
       - id: ruff-format
   - repo: https://github.com/PyCQA/flake8
-    rev: 7.3.0
+    rev: 7.4.1
     hooks:
       - id: flake8
         additional_dependencies:
@@ -37,12 +36,12 @@ repos:
       - id: end-of-file-fixer
       - id: trailing-whitespace
   - repo: https://github.com/asottile/pyupgrade
-    rev: v3.21.2
+    rev: v3.22.0
     hooks:
       - id: pyupgrade
-        args: [--py311-plus]
+        args: [--py312-plus]
   - repo: https://github.com/adrienverge/yamllint.git
-    rev: v1.37.1
+    rev: v1.38.0
     hooks:
       - id: yamllint
         exclude: ^(\.clang-format|\.clang-tidy)$
@@ -59,13 +58,6 @@ repos:
         language: system
         types: [python]
         files: ^esphome/.+\.py$
-      - id: clang-tidy-hash
-        name: Update clang-tidy hash
-        entry: python script/clang_tidy_hash.py --update-if-changed
-        language: python
-        files: ^(\.clang-tidy|platformio\.ini|requirements_dev\.txt)$
-        pass_filenames: false
-        additional_dependencies: []
       - id: ci-custom
         name: ci-custom
         entry: python script/run-in-env.py script/ci-custom.py

AGENTS.md

@@ -9,12 +9,12 @@ This document provides essential context for AI models interacting with this pro
 
 ## 2. Core Technologies & Stack
 
-*   **Languages:** Python (>=3.11), C++ (gnu++20)
+*   **Languages:** Python (>=3.12), C++ (gnu++20)
 *   **Frameworks & Runtimes:** PlatformIO, Arduino, ESP-IDF.
 *   **Build Systems:** PlatformIO is the primary build system. CMake is used as an alternative.
 *   **Configuration:** YAML.
 *   **Key Libraries/Dependencies:**
-    *   **Python:** `voluptuous` (for configuration validation), `PyYAML` (for parsing configuration files), `paho-mqtt` (for MQTT communication), `tornado` (for the web server), `aioesphomeapi` (for the native API).
+    *   **Python:** `voluptuous` (for configuration validation), `PyYAML` (for parsing configuration files), `paho-mqtt` (for MQTT communication), `aioesphomeapi` (for the native API).
     *   **C++:** `ArduinoJson` (for JSON serialization/deserialization), `AsyncMqttClient-esphome` (for MQTT), `ESPAsyncWebServer` (for the web server).
 *   **Package Manager(s):** `pip` (for Python dependencies), `platformio` (for C++/PlatformIO dependencies).
 *   **Communication Protocols:** Protobuf (for native API), MQTT, HTTP.
@@ -35,7 +35,6 @@ This document provides essential context for AI models interacting with this pro
     2.  **Code Generation** (`esphome/codegen.py`, `esphome/cpp_generator.py`): Manages Python to C++ code generation, template processing, and build flag management.
     3.  **Component System** (`esphome/components/`): Contains modular hardware and software components with platform-specific implementations and dependency management.
     4.  **Core Framework** (`esphome/core/`): Manages the application lifecycle, hardware abstraction, and component registration.
-    5.  **Dashboard** (`esphome/dashboard/`): A web-based interface for device configuration, management, and OTA updates.
 
 *   **Platform Support:**
     1.  **ESP32** (`components/esp32/`): Espressif ESP32 family. Supports multiple variants (Original, C2, C3, C5, C6, H2, P4, S2, S3) with ESP-IDF framework. Arduino framework supports only a subset of the variants (Original, C3, S2, S3).
@@ -45,6 +44,16 @@ This document provides essential context for AI models interacting with this pro
 
 ## 4. Coding Conventions & Style Guide
 
+**Read the developer documentation before writing a component.** https://developers.esphome.io covers the
+component lifecycle, the main loop, and the reasoning behind the rules below in far more depth than this
+file does, and it is the authority when they disagree. The most useful starting points:
+
+*   https://developers.esphome.io/architecture/components/ - component lifecycle, `setup()`, `loop()`,
+    setup priorities, and how a component is registered.
+*   https://developers.esphome.io/architecture/components/advanced/ - choosing between `loop()`,
+    `set_interval`, `set_timeout` and `defer`; waking the loop from another thread; the RAM cost of each.
+*   https://developers.esphome.io/contributing/code/ - contribution rules, public API and breaking changes.
+
 *   **Formatting:**
     *   **Python:** Uses `ruff` and `flake8` for linting and formatting. Configuration is in `pyproject.toml`.
     *   **C++:** Uses `clang-format` for formatting. Configuration is in `.clang-format`.
@@ -58,6 +67,26 @@ This document provides essential context for AI models interacting with this pro
         - Function-local constants: `lower_snake_case`
         - Protected/private fields: `lower_snake_case_with_trailing_underscore_`
         - Favor descriptive names over abbreviations
+        - Device components are named brand plus model (`siemens_ira211`), without a protocol-family prefix
+        - Enumerator names: prefix every value of an `enum class` with the enum name converted to
+          `UPPER_SNAKE_CASE` (e.g. `UARTFlushResult::UART_FLUSH_RESULT_SUCCESS`). Never use bare
+          names like `SUCCESS`, `FAILURE`, `OK`, or `FAIL`: platform SDK headers define macros with
+          these common names (for example the Realtek SDKs used by LibreTiny define
+          `#define SUCCESS 0` in `basic_types.h`), and the preprocessor replaces the enumerator
+          before the compiler sees it, breaking the build and clang-tidy on those platforms.
+
+*   **Python Idioms:**
+    *   **Assignment expressions (PEP 572):** Prefer the walrus operator (`:=`) wherever it removes a redundant lookup or a throwaway temporary. The most common case in component code is presence-checking a config key and then indexing it separately — fetch once with `.get()` and bind in the condition instead:
+        ```python
+        # Bad - looks up CONF_BLAH twice
+        if CONF_BLAH in config:
+            cg.add(var.set_blah(config[CONF_BLAH]))
+
+        # Good - single lookup, value bound inline
+        if (blah := config.get(CONF_BLAH)) is not None:
+            cg.add(var.set_blah(blah))
+        ```
+        The same applies to `while` loops and comprehensions where it avoids recomputing a value. Don't contort code to use it — reach for `:=` only when it genuinely cuts repetition or an extra assignment line.
 
 *   **C++ Field Visibility:**
     *   **Prefer `protected`:** Use `protected` for most class fields to enable extensibility and testing. Fields should be `lower_snake_case_with_trailing_underscore_`.
@@ -115,6 +144,9 @@ This document provides essential context for AI models interacting with this pro
 
 *   **C++ Preprocessor Directives:**
     *   **Avoid `#define` for constants:** Using `#define` for constants is discouraged and should be replaced with `const` variables or enums.
+    *   **A constant with a compile-time initializer is `static constexpr`,** not `static const`, at namespace scope
+        and inside functions; `static const` is only for a value computed at run time. A function-local
+        `static constexpr` table is a static constant to clang-tidy, so it is `UPPER_SNAKE_CASE` too.
     *   **Use `#define` only for:**
         - Conditional compilation (`#ifdef`, `#ifndef`)
         - Compile-time sizes calculated during Python code generation (e.g., configuring `std::array` or `StaticVector` dimensions via `cg.add_define()`)
@@ -124,6 +156,86 @@ This document provides essential context for AI models interacting with this pro
     *   **Indentation:** Use spaces (two per indentation level), not tabs
     *   **Type aliases:** Prefer `using type_t = int;` over `typedef int type_t;`
     *   **Line length:** Wrap lines at no more than 120 characters
+    *   **Timing in `loop()`:** Never call `millis()` in a `loop()` body. The current tick's timestamp is
+        already cached - use `App.get_loop_component_start_time()` (from `esphome/core/application.h`).
+        Only reach for `millis()` when you genuinely need sub-tick resolution inside a long operation.
+    *   **The main loop runs every 16 ms.** A rate-limit gate shorter than that does nothing: the check
+        passes on essentially every pass of the loop, so it costs a comparison and buys nothing. Pick an
+        interval comfortably coarser than 16 ms, or drop the gate entirely and accept running every loop.
+        ```cpp
+        // Bad - a 10ms gate against a 16ms loop never holds anything back
+        static constexpr uint32_t POLL_INTERVAL_MS = 10;
+        const uint32_t now = millis();
+        if (now - this->last_poll_ < POLL_INTERVAL_MS)
+          return;
+        this->last_poll_ = now;
+        ```
+        ```cpp
+        // Good - an interval that actually rate limits, off the cached timestamp
+        static constexpr uint32_t POLL_INTERVAL_MS = 100;
+        const uint32_t now = App.get_loop_component_start_time();
+        if (now - this->last_poll_ < POLL_INTERVAL_MS)
+          return;
+        this->last_poll_ = now;
+        ```
+        Pick the primitive by cadence: under 250 ms use a gated `loop()`; 500 ms and above use
+        `set_interval`. Full reasoning, including why `set_interval` costs more below 500 ms:
+        https://developers.esphome.io/architecture/components/advanced/#quick-rule-of-thumb
+    *   **Repeating work:** use `set_interval`, never a `set_timeout` that re-arms itself; every re-arm
+        rebuilds the callback and re-queues a scheduler item, and the period drifts.
+    *   **Scheduler ids:** name a timer only when it must be cancelled or replaced, and use a
+        `static constexpr uint32_t` id, never a string. Ids are per component instance and cannot clash with other
+        components, so number them from 0 and keep all of a component's ids together in one place.
+        ```cpp
+        static constexpr uint32_t READ_TIMEOUT_ID = 0;
+        this->set_timeout(READ_TIMEOUT_ID, 50, [this]() { this->read_(); });
+        ```
+    *   **Don't override a default with the same value:** if a base class method already returns what you
+        want, do not override it. `Component::get_setup_priority()` returns `setup_priority::DATA`, so a
+        component that wants `DATA` should simply leave it alone.
+        ```cpp
+        // Bad - this is exactly what the base class already does
+        float get_setup_priority() const override { return setup_priority::DATA; }
+        ```
+    *   **Constant data lives in RAM on ESP8266** unless marked `PROGMEM`: `const` tables, class-scope
+        constant arrays, and in-class `std::array` initializers (a `.rodata` template plus a copy). Mark the
+        table `PROGMEM` and read it with `progmem_memcpy()` or `progmem_read_byte()` (`esphome/core/hal.h`);
+        for a table of structs copy the whole entry out and return the copy, never a pointer into flash.
+        On other platforms they are a plain `memcpy` and dereference, so no `#ifdef` is needed. Declare log tags with
+        `ESPHOME_LOG_TAG(TAG, "name")`, pass status text as `LOG_STR`, e.g.
+        `this->mark_failed(LOG_STR(ESP_LOG_MSG_COMM_FAIL))`, and give an action whose payload is a byte array
+        a `TEMPLATABLE_BYTES` field so a static code stays in flash.
+    *   **Fixed sets of names** (status words, model names): a `PROGMEM_STRING_TABLE` (`esphome/core/progmem.h`),
+        logged with `get_log_str` and `LOG_STR_ARG`, copied out with `ESPHOME_strncpy_P` before `publish_state`.
+        A `dump_config` helper that picks a name returns `const LogString *` from `LOG_STR(...)`.
+    *   **CRCs and bit tricks** come from `esphome/core/helpers.h` (`crc8`, `crc16`, `reverse_bits`, `encode_uint16`);
+        pick the helper's polynomial and bit-order arguments instead of writing a CRC loop.
+    *   **An unset float is `NAN`**, not `optional<float>`; the optional costs 4 to 8 bytes of padding per member.
+    *   **`PollingComponent` first runs `update()`** after a random share of the interval; call `this->update()` at the
+        end of `setup()` when the entities need a state at boot.
+    *   **A multi-step command sequence** leaves the step pending when the answer is malformed, so the existing
+        resend and give-up logic handles it; never advance on an answer that was only logged.
+    *   **Entities that each watch a register field:** one class taking register, shift, mask and match from a Python
+        tuple table, in typed `StaticVector`s sized by `cg.slot_counter`; not a template per bit, not a virtual
+        listener base (a second vptr per entity).
+    *   **A global feature define** says the code is compiled in, not that this instance uses it: keep the
+        per-instance check (`if (this->aux_device_ == ...)`) inside the `#ifdef`.
+    *   **Model-specific code** sits behind a define the Python side emits only for configs that use that model
+        (`cg.add_define("USE_SELEC_METER_EM4M")`), so the other models' builds leave it out; the runtime model check
+        stays inside the `#ifdef`.
+    *   **Failures after `setup()`:** `mark_failed()` is for `setup()` only. A runtime read or write failure
+        sets `status_set_warning()` and the next success clears it; writes skip work when `is_failed()`.
+    *   **Fixed platform facts compile to nothing:** gate on `#ifdef USE_ESP8266` and friends, not on a
+        runtime check of something the build already knows.
+    *   **Logging string literals:** wrap literals passed as `%s` arguments in `LOG_STR_LITERAL()` so they
+        can be stored in flash rather than RAM.
+        ```cpp
+        // Bad
+        ESP_LOGV(TAG, "Key %u %s", key, pressed ? "pressed" : "released");
+
+        // Good
+        ESP_LOGV(TAG, "Key %u %s", key, pressed ? LOG_STR_LITERAL("pressed") : LOG_STR_LITERAL("released"));
+        ```
     *   **Constructor parameters vs setters:** Component properties that are both **required** and **invariant**
         (never change after construction) should be constructor parameters rather than set via setter methods.
         This makes the dependency explicit and prevents use of the object in an incompletely-initialized state.
@@ -161,6 +273,10 @@ This document provides essential context for AI models interacting with this pro
         ```
 
     *   **Component Metadata:**
+        - `DOMAIN`: Required in every component's `__init__.py`, set to the component's own name as a plain
+          string (e.g. `DOMAIN = "my_component"`). Other code refers to the component through it, e.g.
+          `from esphome.components.my_component import DOMAIN`. Keep it in alphabetical order with the other
+          metadata constants. CI (`lint_component_domain`) fails without it.
         - `DEPENDENCIES`: List of required components
         - `AUTO_LOAD`: Components to automatically load
         - `CONFLICTS_WITH`: Incompatible components
@@ -174,16 +290,20 @@ This document provides essential context for AI models interacting with this pro
         import esphome.config_validation as cv
         from esphome.const import CONF_KEY, CONF_ID
 
+        DOMAIN = "my_component"
         CONF_PARAM = "param"  # A constant that does not yet exist in esphome/const.py
 
         my_component_ns = cg.esphome_ns.namespace("my_component")
         MyComponent = my_component_ns.class_("MyComponent", cg.Component)
 
-        CONFIG_SCHEMA = cv.Schema({
-            cv.GenerateID(): cv.declare_id(MyComponent),
-            cv.Required(CONF_KEY): cv.string,
-            cv.Optional(CONF_PARAM, default=42): cv.int_,
-        }).extend(cv.COMPONENT_SCHEMA)
+        CONFIG_SCHEMA = cv.Schema(
+            {
+                cv.GenerateID(): cv.declare_id(MyComponent),
+                cv.Required(CONF_KEY): cv.string,
+                cv.Optional(CONF_PARAM, default=42): cv.int_,
+            }
+        ).extend(cv.COMPONENT_SCHEMA)
+
 
         async def to_code(config):
             var = cg.new_Pvariable(config[CONF_ID])
@@ -217,7 +337,12 @@ This document provides essential context for AI models interacting with this pro
         - **Sensor:**
           ```python
           from esphome.components import sensor
-          CONFIG_SCHEMA = sensor.sensor_schema(MySensor).extend(cv.polling_component_schema("60s"))
+
+          CONFIG_SCHEMA = sensor.sensor_schema(MySensor).extend(
+              cv.polling_component_schema("60s")
+          )
+
+
           async def to_code(config):
               var = await sensor.new_sensor(config)
               await cg.register_component(var, config)
@@ -226,7 +351,10 @@ This document provides essential context for AI models interacting with this pro
         - **Binary Sensor:**
           ```python
           from esphome.components import binary_sensor
-          CONFIG_SCHEMA = binary_sensor.binary_sensor_schema().extend({ ... })
+
+          CONFIG_SCHEMA = binary_sensor.binary_sensor_schema().extend({...})
+
+
           async def to_code(config):
               var = await binary_sensor.new_binary_sensor(config)
           ```
@@ -234,11 +362,33 @@ This document provides essential context for AI models interacting with this pro
         - **Switch:**
           ```python
           from esphome.components import switch
-          CONFIG_SCHEMA = switch.switch_schema().extend({ ... })
+
+          CONFIG_SCHEMA = switch.switch_schema().extend({...})
+
+
           async def to_code(config):
               var = await switch.new_switch(config)
           ```
 
+        - **Optional child entities of a hub:** bind the config once with `sensor.sub_sensors(config)` (or
+          `sub_binary_sensors`, `sub_text_sensors`, `sub_buttons`, `sub_switches`, `sub_numbers`,
+          `sub_selects` in their domains), adding `parent=hub` for entities that derive from `Parented<T>`,
+          then make one call per key, even when there is only one. A call creates the entity only when its key
+          is configured, passes it to the setter and returns it (or `None`); extra arguments such as
+          `min_value` or `options` go on the call. Always name the setter explicitly on the object that owns
+          it, never with `getattr` and an f-string, and keep that variable short (`var` for the component
+          itself, `hub` for one fetched with `cg.get_variable`) so the calls fit on one line. Loops whose
+          setter also takes an index, such as `set_gate_threshold(x, n)`, stay as they are.
+          ```python
+          async def to_code(config):
+              var = cg.new_Pvariable(config[CONF_ID])
+              sensors = sensor.sub_sensors(config)
+              await sensors(CONF_TEMPERATURE, var.set_temperature_sensor)
+              await sensors(CONF_HUMIDITY, var.set_humidity_sensor)
+              buttons = button.sub_buttons(config, parent=var)
+              await buttons(CONF_RESTART, var.set_restart_button)
+          ```
+
 *   **Automations (Triggers, Actions, Conditions):**
 
     Automations have three building blocks: **Triggers** (fire when something happens), **Actions** (do something), and **Conditions** (check if something is true).
@@ -251,10 +401,13 @@ This document provides essential context for AI models interacting with this pro
         ```python
         from esphome import automation
 
-        CONFIG_SCHEMA = cv.Schema({
-            cv.GenerateID(): cv.declare_id(MyComponent),
-            cv.Optional(CONF_ON_STATE): automation.validate_automation({}),
-        }).extend(cv.COMPONENT_SCHEMA)
+        CONFIG_SCHEMA = cv.Schema(
+            {
+                cv.GenerateID(): cv.declare_id(MyComponent),
+                cv.Optional(CONF_ON_STATE): automation.validate_automation({}),
+            }
+        ).extend(cv.COMPONENT_SCHEMA)
+
 
         async def to_code(config):
             var = cg.new_Pvariable(config[CONF_ID])
@@ -279,6 +432,25 @@ This document provides essential context for AI models interacting with this pro
                 )
         ```
 
+        When the callback's parameters differ from the automation's arguments, or the trigger should only
+        fire for some values, pass `params`, `forward` and `when`; the helper then generates a capture-less
+        lambda (stored inline, nothing allocated). Name the parent with `automation.parent_ref(var)` and build
+        `forward` from `MockObj` calls, not f-strings of C++:
+        ```python
+        parent = automation.parent_ref(var)
+        index = cg.RawExpression("index")
+        await automation.build_callback_automation(
+            var,
+            "add_on_state_callback",
+            [(cg.StringRef, "x"), (cg.size_t, "i")],
+            conf,
+            params=[(cg.size_t, "index")],
+            forward=[cg.StringRef(parent.option_at(index)), index],
+        )
+        ```
+        `build_parent_callback_automation`, `build_trigger_callback`, `build_callback_automations` and
+        `build_trigger_automations` cover the other shapes; see their docstrings in `esphome/automation.py`.
+
         **C++ -- no trigger class needed.** The callback registration method must be templatized to accept both `std::function` and lightweight forwarder structs (which avoid heap allocation):
         ```cpp
         class MyComponent : public Component {
@@ -304,11 +476,14 @@ This document provides essential context for AI models interacting with this pro
         ```python
         TurnOnTrigger = my_ns.class_("TurnOnTrigger", automation.Trigger.template())
 
-        CONFIG_SCHEMA = cv.Schema({
-            cv.Optional(CONF_ON_TURN_ON): automation.validate_automation(
-                {cv.GenerateID(CONF_TRIGGER_ID): cv.declare_id(TurnOnTrigger)}
-            ),
-        })
+        CONFIG_SCHEMA = cv.Schema(
+            {
+                cv.Optional(CONF_ON_TURN_ON): automation.validate_automation(
+                    {cv.GenerateID(CONF_TRIGGER_ID): cv.declare_id(TurnOnTrigger)}
+                ),
+            }
+        )
+
 
         async def to_code(config):
             for conf in config.get(CONF_ON_TURN_ON, []):
@@ -342,7 +517,31 @@ This document provides essential context for AI models interacting with this pro
           MyComponent *parent_;
         };
         ```
-        Register with `@automation.register_action("my_component.do_something", MyAction, schema, synchronous=True)`. Use `synchronous=True` for actions that run to completion inside `play()` without deferring. Use `synchronous=False` if the action may suspend/defer execution (e.g. `delay`, `wait_until`, `script.wait`) or store trigger arguments for later use.
+        Register it without writing a builder:
+        ```python
+        automation.register_simple_action(
+            "my_component.do_something", MyAction, schema, synchronous=True
+        )
+        ```
+        The constructor receives the object named by `config[CONF_ID]`. Use `register_bare_action` for a
+        no-argument constructor, `register_parented_action` for a class deriving from `Parented<T>`, and
+        the `@automation.register_action(...)` decorator only when the builder must also set fields.
+
+        Use `synchronous=True` for actions that run to completion inside `play()` without deferring. Use `synchronous=False` if the action may suspend/defer execution (e.g. `delay`, `wait_until`, `script.wait`) or store trigger arguments for later use.
+
+        **Actions that only forward templatable values to their parent need no C++ class.** Register them
+        with `register_apply_action`; do not write a `TEMPLATABLE_VALUE` class or a builder for this shape.
+        ```python
+        automation.register_apply_action(
+            "my_component.set_gains",
+            schema,
+            automation.ApplyField(CONF_KP, "set_kp", cg.float_),
+            automation.ApplyField(CONF_KI, "set_ki", cg.float_),
+        )
+        ```
+        The `ApplyField`, `ApplyCall` and `register_apply_action` docstrings in `esphome/automation.py` cover
+        the rest; `cover.control` and `cover.template.publish` are in-tree examples. `TEMPLATABLE_VALUE` with
+        `cg.templatable` stays for actions whose `play()` has real logic beyond forwarding values.
 
     *   **Conditions:**
         ```cpp
@@ -354,9 +553,30 @@ This document provides essential context for AI models interacting with this pro
           MyComponent *parent_;
         };
         ```
-        Register with `@automation.register_condition("my_component.is_active", MyCondition, schema)`.
+        Register with `automation.register_simple_condition("my_component.is_active", MyCondition, schema)`;
+        `register_bare_condition`, `register_parented_condition` and the decorator follow the action rules.
+
+        **Conditions that only test their parent need no C++ class either.** Register them with
+        `register_apply_condition`; the expression is applied to the parent, and an `ApplyCall` compares
+        against config values.
+        ```python
+        automation.register_apply_condition("my_component.is_active", schema, "is_active()")
+        automation.register_apply_condition(
+            "my_component.state_is",
+            schema,
+            automation.ApplyCall("state == {}", ((CONF_STATE, cg.bool_),)),
+        )
+        ```
+        `cover.is_open`, `rtttl.is_playing` and `component.is_idle` are in-tree examples.
+
+*   **Type Hints:** Type-hint all function signatures, including test functions and config validators (e.g. `def validate_x(config: ConfigType) -> ConfigType:`, `def test_x() -> None:`). Import `ConfigType` from `esphome.types`.
+    An intentionally empty body (an unused `register_dumper` builder, for example) gets a one-line docstring, not `pass`.
 
 *   **Configuration Validation:**
+    *   **One schema per model:** when a component's options depend on a `model` key, use `cv.typed_schema({...},
+        key=CONF_MODEL, default_type=...)` with each model's own keys and defaults, not one schema plus a validator
+        that rejects keys and fills defaults afterwards.
+    *   **Reuse existing validators:** Before writing a custom validator, check for an existing one in `config_validation.py` and compose it in `cv.All(...)` rather than duplicating logic across components. For example, rename a config key with `cv.rename_key(CONF_OLD, CONF_NEW, removed_in="2026.6.0")`, and reject mutually-exclusive keys with `cv.has_at_most_one_key(...)` / `cv.has_exactly_one_key(...)`. See how `api` composes `cv.has_exactly_one_key` + `cv.rename_key`.
     *   **Common Validators:** `cv.int_`, `cv.float_`, `cv.string`, `cv.boolean`, `cv.int_range(min=0, max=100)`, `cv.positive_int`, `cv.percentage`.
     *   **Complex Validation:** `cv.All(cv.string, cv.Length(min=1, max=50))`, `cv.Any(cv.int_, cv.string)`.
     *   **Platform-Specific:** `cv.only_on(["esp32", "esp8266"])`, `esp32.only_on_variant(...)`, `cv.only_on_esp32`, `cv.only_on_esp8266`, `cv.only_on_rp2040`.
@@ -369,6 +589,7 @@ This document provides essential context for AI models interacting with this pro
          .extend(i2c.i2c_device_schema(0x48))
          .extend(spi.spi_device_schema(cs_pin_required=True))
         ```
+    *   **Constants:** `esphome/const.py` is frozen — do not add new `CONF_` constants there. Define a component-local constant in the component's own `.py` (as with `CONF_PARAM` above); for a constant shared by multiple components, add it to `esphome/components/const/__init__.py`. CI (`lint_constants_usage`) fails if the same constant is defined in three or more component files. Constants used in core files (i.e. those not under `esphome/components`) may be added to `esphome/const.py` but will require adjustment to the CI validation check.
 
 ## 5. Key Files & Entrypoints
 
@@ -376,7 +597,7 @@ This document provides essential context for AI models interacting with this pro
 *   **Configuration:**
     *   `pyproject.toml`: Defines the Python project metadata and dependencies.
     *   `platformio.ini`: Configures the PlatformIO build environments for different microcontrollers.
-    *   `.pre-commit-config.yaml`: Configures the pre-commit hooks for linting and formatting.
+    *   `.pre-commit-config.yaml`: Configures the lint and format hooks, run by `prek`.
 *   **CI/CD Pipeline:** Defined in `.github/workflows`.
 *   **Static Analysis & Development:**
     *   `esphome/core/defines.h`: A comprehensive header file containing all `#define` directives that can be added by components using `cg.add_define()` in Python. This file is used exclusively for development, static analysis tools, and CI testing - it is not used during runtime compilation. When developing components that add new defines, they must be added to this file to ensure proper IDE support and static analysis coverage. The file includes feature flags, build configurations, and platform-specific defines that help static analyzers understand the complete codebase without needing to compile for specific platforms.
@@ -384,10 +605,22 @@ This document provides essential context for AI models interacting with this pro
 ## 6. Development & Testing Workflow
 
 *   **Local Development Environment:** Use the provided Docker container or create a Python virtual environment and install dependencies from `requirements_dev.txt`.
-*   **Running Commands:** Use the `script/run-in-env.py` script to execute commands within the project's virtual environment. For example, to run the linter: `python3 script/run-in-env.py pre-commit run`.
+*   **Running Commands:** Use the `script/run-in-env.py` script to execute commands within the project's virtual environment. For example, to run the linter: `python3 script/run-in-env.py prek run`.
 *   **Testing:**
     *   **Python:** Run unit tests with `pytest`.
     *   **C++:** Use `clang-tidy` for static analysis.
+    *   **C++ unit tests:** gtests in `tests/components/<component>/*.cpp`, run with
+        `script/cpp_unit_test.py <component>`; see `tests/components/README.md`. Set the loop clock in a test
+        with `LoopBlockingGuard{nullptr, nullptr, now_ms}`; the runner resets it before every test.
+    *   **Integration tests:** `tests/integration/` runs real firmware on the host. Serial components prove
+        timeouts and recovery with the `uart_mock` fixture (`tests/integration/test_uart_mock_*.py`). Its
+        `responses` match on content every time they see it; to answer a repeated request differently, use an
+        `on_tx` lambda with a static counter and `inject_to_rx_buffer` (see `uart_mock_ld2412_requery_after_giveup.yaml`).
+    *   **Keep build-test variants few:** one `test.<platform>.yaml` per platform the code actually differs
+        on. A variant per board for platform-identical code only adds CI time.
+    *   **A test that needs a shared component in a particular state** (`esp32_ble` security settings, an option
+        that PSRAM rejects) cannot be grouped: the merged block wins and the whole batch fails. Add the component to
+        `ISOLATED_COMPONENTS` in `script/analyze_component_buses.py`, or move the option to a `test-<variant>` file.
     *   **Component Tests:** YAML-based compilation tests are located in `tests/`. The structure is as follows:
         ```
         tests/
@@ -415,13 +648,14 @@ This document provides essential context for AI models interacting with this pro
 
         When a PR's only edits to a component are `validate.*.yaml` files (no source changes, no `test.*.yaml` changes, and the component isn't pulled in as a dependency of another changed component), CI skips the compile stage for that component entirely and only runs config validation. This is decided in `script/determine-jobs.py` via `_component_change_is_validate_only` and surfaced as the `validate_only_components` output that the `test-build-components-split` job consumes.
 
-    *   **Test Grouping with Packages:** Components that use shared bus packages can be grouped together in CI to reduce build count. **Never define buses (uart, i2c, spi, modbus) directly in test YAML files** — always use packages from `test_build_components/common/`:
+    *   **Test Grouping with Packages:** Components that use shared bus packages can be grouped together in CI to reduce build count. **Never define buses (uart, i2c, spi, modbus) directly in test YAML files** — always use packages from `test_build_components/common/`.
+
+        All includes in test files must go through dict-style `packages:` so that batch grouping works correctly — the grouping scripts only understand dict-style packages. Never use list-style packages (`packages: [- !include ...]`) or top-level merge keys (`<<: !include common.yaml`). Bus packages are keyed by the bus name; the component's `common.yaml` is keyed by the component name (e.g. `cst328: !include common.yaml`):
         ```yaml
-        # test.esp32-idf.yaml — use packages for buses
+        # test.esp32-idf.yaml — everything included via named packages
         packages:
           uart: !include ../../test_build_components/common/uart_115200/esp32-idf.yaml
-
-        <<: !include common.yaml
+          my_component: !include common.yaml
         ```
         ```yaml
         # common.yaml — component config only, NO bus definitions
@@ -433,6 +667,7 @@ This document provides essential context for AI models interacting with this pro
             name: My Sensor
         ```
         Components that define buses directly are flagged as "NEEDS MIGRATION" and cannot be grouped, increasing CI build time.
+        `!extend` and `!remove` in a `test.*.yaml` also make the component ungroupable; add a standalone entity with its own id instead.
 
     *   **Testing All Components Together:** To verify that all components can be tested together without ID conflicts or configuration issues, use:
         ```bash
@@ -443,7 +678,6 @@ This document provides essential context for AI models interacting with this pro
     *   **Debug Tools:**
         - `esphome config <file>.yaml` to validate configuration.
         - `esphome compile <file>.yaml` to compile without uploading.
-        - Check the Dashboard for real-time logs.
         - Use component-specific debug logging.
     *   **Common Issues:**
         - **Import Errors**: Check component dependencies and `PYTHONPATH`.
@@ -457,9 +691,10 @@ This document provides essential context for AI models interacting with this pro
     1.  **Fork & Branch:** Create a new branch based on the `dev` branch (always use `git checkout -b <branch-name> dev` to ensure you're branching from `dev`, not the currently checked out branch).
     2.  **Make Changes:** Adhere to all coding conventions and patterns.
     3.  **Test:** Create component tests for all supported platforms and run the full test suite locally.
-    4.  **Lint:** Run `pre-commit` to ensure code is compliant.
+    4.  **Lint:** Run `prek` to ensure code is compliant.
     5.  **Commit:** Commit your changes. There is no strict format for commit messages.
-    6.  **Pull Request:** Submit a PR against the `dev` branch. The Pull Request title should have a prefix of the component being worked on (e.g., `[display] Fix bug`, `[abc123] Add new component`). Update documentation, examples, and add `CODEOWNERS` entries as needed. Pull requests should always be made using the `.github/PULL_REQUEST_TEMPLATE.md` template - fill out all sections completely without removing any parts of the template.
+    6.  **Pull Request:** Submit a PR against the `dev` branch. The Pull Request title must start with a `[tag]` prefix. For component work, use the component name (e.g., `[display] Fix bug`, `[abc123] Add new component`); for changes to shared/core code that isn't tied to a single component, use `[core]` (e.g., `[core] Add validator`). Update documentation, examples, and add `CODEOWNERS` entries as needed. Pull requests should always be made using the `.github/PULL_REQUEST_TEMPLATE.md` template - fill out all sections completely without removing any parts of the template.
+    7.  **Comments:** When commenting on GitHub PRs or issues, don't tag contributors, especially bots. Avoid referring to list items (e.g. from reviews) with the form #nn - this will be interpreted by GitHub as a reference to issue or PR nn. Keep comments short and exclude irrelevant details, backstories, restatement of previous comments and anything that is already obvious to the reader.
 
 *   **Documentation Contributions:**
     *   Documentation is hosted in the separate `esphome/esphome.io` repository.
@@ -520,6 +755,36 @@ This document provides essential context for AI models interacting with this pro
            Use `cg.add_define("MAX_SERVICES", count)` to set the size from Python configuration.
            Like `std::array` but with vector-like API (`push_back()`, `size()`) and no STL reallocation code.
 
+           **Listener and child-entity registration lists are the most common case, and the most commonly
+           missed.** A `register_*()` method called once per child at code generation time has a count that
+           is known at compile time, so it should never be a `std::vector`. Use `cg.slot_counter()`: it
+           returns a function that each consumer calls once per slot it will occupy, and after every
+           `to_code` has run it emits the define with the final count. When nothing registers, no define is
+           emitted and the storage plus its registration method compile out entirely.
+           ```python
+           # hub component's __init__.py
+           _request_listener_slot = cg.slot_counter("MY_COMPONENT_LISTENER_COUNT")
+
+
+           async def register_listener(hub: MockObj, var: MockObj) -> None:
+               _request_listener_slot()
+               cg.add(hub.register_listener(var))
+           ```
+           When several instances each own a list declared at the same size (one per hub of a
+           `MULTI_CONF` component), pass the owning object as the key, `_request_listener_slot(str(hub))`;
+           the define is then the largest count any one key requested instead of the total.
+           ```cpp
+           #ifdef MY_COMPONENT_LISTENER_COUNT
+             void register_listener(MyComponentListener *listener);
+           #endif
+            protected:
+           #ifdef MY_COMPONENT_LISTENER_COUNT
+             StaticVector<MyComponentListener *, MY_COMPONENT_LISTENER_COUNT> listeners_;
+           #endif
+           ```
+           Request slots from `to_code`, not from a job that runs after `CoroPriority.FINAL` - a late
+           request raises rather than silently undercounting.
+
         3. **Runtime-known sizes:** Use `FixedVector` from `esphome/core/helpers.h` when the size is only known at runtime initialization.
            ```cpp
            // Bad - generates STL realloc code (_M_realloc_insert)
@@ -557,9 +822,27 @@ This document provides essential context for AI models interacting with this pro
            ```
            Linear search on small datasets (1-16 elements) is often faster than hashing/tree overhead, but this depends on lookup frequency and access patterns. For frequent lookups in hot code paths, the O(1) vs O(n) complexity difference may still matter even for small datasets. `std::vector` with simple structs is usually fine—it's the heavy containers (`map`, `set`, `unordered_map`) that should be avoided for small datasets unless profiling shows otherwise.
 
-        5. **Avoid `std::deque`:** It allocates in 512-byte blocks regardless of element size, guaranteeing at least 512 bytes of RAM usage immediately. This is a major source of crashes on memory-constrained devices.
+        5. **Strings set once from configuration:** Use `StringRef` (`esphome/core/string_ref.h`) rather than
+           `std::string`. Code generation passes a string literal that lives in flash for the life of the
+           program, so storing a `std::string` copies it onto the heap for nothing. `StringRef` is a
+           non-owning pointer plus length; it does not copy, and it must only ever refer to storage that
+           outlives it (a string literal, or a buffer owned elsewhere).
+           ```cpp
+           // Bad - heap copy of a literal that is already in flash
+           void set_keys(std::string keys) { this->keys_ = std::move(keys); }
+           std::string keys_;
+           ```
+           ```cpp
+           // Good - no allocation
+           void set_keys(const char *keys) { this->keys_ = StringRef(keys); }
+           StringRef keys_;
+           ```
 
-        6. **Detection:** Look for these patterns in compiler output:
+        6. **Avoid `std::deque`:** It allocates in 512-byte blocks regardless of element size, guaranteeing at least 512 bytes of RAM usage immediately. This is a major source of crashes on memory-constrained devices.
+
+        7. **Never use `new (std::nothrow)`:** On ESP-IDF exceptions are disabled, so a failed nothrow allocation aborts instead of returning `nullptr`. Use `RAMAllocator` from `esphome/core/helpers.h`; CI rejects `std::nothrow`.
+
+        8. **Detection:** Look for these patterns in compiler output:
            - Large code sections with STL symbols (vector, map, set)
            - `alloc`, `realloc`, `dealloc` in symbol names
            - `_M_realloc_insert`, `_M_default_append` (vector reallocation)
@@ -605,6 +888,7 @@ This document provides essential context for AI models interacting with this pro
         _component_state = []
         _use_feature = None
 
+
         def enable_feature():
             global _use_feature
             _use_feature = True
@@ -624,20 +908,24 @@ This document provides essential context for AI models interacting with this pro
 
         DOMAIN = "my_component"
 
+
         @dataclass
         class MyComponentData:
             feature_enabled: bool = False
             item_count: int = 0
             items: list[str] = field(default_factory=list)
 
+
         def _get_data() -> MyComponentData:
             if DOMAIN not in CORE.data:
                 CORE.data[DOMAIN] = MyComponentData()
             return CORE.data[DOMAIN]
 
+
         def request_feature() -> None:
             _get_data().feature_enabled = True
 
+
         def add_item(item: str) -> None:
             _get_data().items.append(item)
         ```
@@ -645,7 +933,7 @@ This document provides essential context for AI models interacting with this pro
         If you need a real-world example, search for components that use `@dataclass` with `CORE.data` in the codebase. Note: Some components may use `TypedDict` for dictionary-based storage; both patterns are acceptable depending on your needs.
 
         **Why this matters:**
-        - Module-level globals persist between compilation runs if the dashboard doesn't fork/exec
+        - Module-level globals persist between compilation runs if the host process (e.g. device-builder) doesn't fork/exec
         - `CORE.data` automatically clears between runs
         - Namespacing under `DOMAIN` prevents key collisions between components
         - `@dataclass` provides type safety and cleaner attribute access
@@ -692,9 +980,37 @@ This document provides essential context for AI models interacting with this pro
     ```
 
 *   **Deprecation Pattern (Python):**
+    For a renamed config key, use the shared `cv.rename_key` validator with `removed_in` (and `component` for context) — it warns and auto-migrates:
+    ```python
+    CONFIG_SCHEMA = cv.All(
+        cv.rename_key(
+            CONF_OLD_KEY, CONF_NEW_KEY, removed_in="2026.6.0", component="my_component"
+        ),
+        cv.Schema({...}),
+    )
+    ```
+    For other deprecations, warn manually during validation:
     ```python
     # Remove before 2026.6.0
     if CONF_OLD_KEY in config:
-        _LOGGER.warning(f"'{CONF_OLD_KEY}' deprecated, use '{CONF_NEW_KEY}'. Removed in 2026.6.0")
+        _LOGGER.warning(
+            f"'{CONF_OLD_KEY}' deprecated, use '{CONF_NEW_KEY}'. Removed in 2026.6.0"
+        )
         config[CONF_NEW_KEY] = config.pop(CONF_OLD_KEY)  # Auto-migrate
     ```
+## 9. English Language
+
+The project uses English for non-code content. When drafting documentation, code comments, commit messages,
+PR descriptions, and similar text, avoid technical jargon. Instead, express concepts in plain English,
+using standard technical terms only when required. Ensure the text is readily comprehensible to a wide
+audience, including non-native English speakers.
+
+## 10. Code Comments
+
+Code comments on individual lines should be used only where necessary to flag issues that may not be obvious
+on a simple reading of the code. Keep them short (e.g. 1 or 2 lines).
+
+Function and method comment blocks may include more detail as required to make
+calling contracts clear and document parameter usage, but should still be kept concise.
+
+Avoid redundancy and repetition; comments should never simply restate what the code already says.

CODEOWNERS

@@ -19,7 +19,6 @@ esphome/components/ac_dimmer/* @glmnet
 esphome/components/adc/* @esphome/core
 esphome/components/adc128s102/* @DeerMaximum
 esphome/components/addressable_light/* @justfalter
-esphome/components/ade7880/* @kpfleming
 esphome/components/ade7953/* @angelnu
 esphome/components/ade7953_base/* @angelnu
 esphome/components/ade7953_i2c/* @angelnu
@@ -28,7 +27,7 @@ esphome/components/ads1118/* @solomondg1
 esphome/components/ags10/* @mak-42
 esphome/components/aic3204/* @kbx81
 esphome/components/airthings_ble/* @jeromelaban
-esphome/components/airthings_wave_base/* @jeromelaban @kpfleming @ncareau
+esphome/components/airthings_wave_base/* @jeromelaban @ncareau
 esphome/components/airthings_wave_mini/* @ncareau
 esphome/components/airthings_wave_plus/* @jeromelaban @precurse
 esphome/components/alarm_control_panel/* @grahambrown11 @hwstar
@@ -40,9 +39,12 @@ esphome/components/am43/sensor/* @buxtronix
 esphome/components/analog_threshold/* @ianchi
 esphome/components/animation/* @syndlex
 esphome/components/anova/* @buxtronix
+esphome/components/apc1/* @ademuri
 esphome/components/apds9306/* @aodrenah
+esphome/components/apds9930/* @iyesin
 esphome/components/api/* @esphome/core
 esphome/components/aqi/* @freekode @jasstrong @ximex
+esphome/components/aranet/* @patrickod
 esphome/components/as5600/* @ammmze
 esphome/components/as5600/sensor/* @ammmze
 esphome/components/as7341/* @mrgnr
@@ -66,16 +68,21 @@ esphome/components/bedjet/climate/* @jhansche
 esphome/components/bedjet/fan/* @jhansche
 esphome/components/bedjet/sensor/* @javawizard @jhansche
 esphome/components/beken_spi_led_strip/* @Mat931
+esphome/components/bh1745/* @latonita
 esphome/components/bh1750/* @OttoWinter
 esphome/components/bh1900nux/* @B48D81EFCC
 esphome/components/binary_sensor/* @esphome/core
 esphome/components/bk72xx/* @kuba2k2
+esphome/components/bk72xx_ble/* @Bl00d-B0b
+esphome/components/bk72xx_ble_tracker/* @Bl00d-B0b
 esphome/components/bl0906/* @athom-tech @jesserockz @tarontop
 esphome/components/bl0939/* @ziceva
 esphome/components/bl0940/* @dan-s-github @tobias-
 esphome/components/bl0942/* @dbuezas @dwmw2
 esphome/components/ble_client/* @buxtronix @clydebarrow
+esphome/components/ble_device_base/* @Bl00d-B0b
 esphome/components/ble_nus/* @tomaszduda23
+esphome/components/bluetooth_connection/* @bdraco @jesserockz
 esphome/components/bluetooth_proxy/* @bdraco @jesserockz
 esphome/components/bm8563/* @abmantis
 esphome/components/bme280_base/* @esphome/core
@@ -84,6 +91,7 @@ esphome/components/bme680_bsec/* @trvrnrth
 esphome/components/bme68x_bsec2/* @kbx81 @neffs
 esphome/components/bme68x_bsec2_i2c/* @kbx81 @neffs
 esphome/components/bmi160/* @flaviut
+esphome/components/bmi270/* @clydebarrow
 esphome/components/bmp280_base/* @ademuri
 esphome/components/bmp280_i2c/* @ademuri
 esphome/components/bmp280_spi/* @ademuri
@@ -96,6 +104,10 @@ esphome/components/bmp581_i2c/* @danielkent-net @kahrendt
 esphome/components/bmp581_spi/* @danielkent-net @kahrendt
 esphome/components/bp1658cj/* @Cossid
 esphome/components/bp5758d/* @Cossid
+esphome/components/bq25186/* @Rapsssito
+esphome/components/bq27220/* @rggammon
+esphome/components/bq76972/* @limpkin
+esphome/components/bridge/* @kbx81
 esphome/components/bthome_mithermometer/* @nagyrobi
 esphome/components/button/* @esphome/core
 esphome/components/bytebuffer/* @clydebarrow
@@ -107,6 +119,8 @@ esphome/components/captive_portal/* @esphome/core
 esphome/components/cc1101/* @gabest11 @lygris
 esphome/components/ccs811/* @habbie
 esphome/components/cd74hc4067/* @asoehlke
+esphome/components/cdc_acm_uart/* @kbx81
+esphome/components/cdc_acm_uart/bridge/* @kbx81
 esphome/components/ch422g/* @clydebarrow @jesterret
 esphome/components/ch423/* @dwmw2
 esphome/components/chsc6x/* @kkosik20
@@ -118,13 +132,17 @@ esphome/components/combination/* @Cat-Ion @kahrendt
 esphome/components/const/* @esphome/core
 esphome/components/coolix/* @glmnet
 esphome/components/copy/* @OttoWinter
+esphome/components/counter/* @clydebarrow
 esphome/components/cover/* @esphome/core
 esphome/components/cs5460a/* @balrog-kun
 esphome/components/cse7761/* @berfenger
 esphome/components/cst226/* @clydebarrow
+esphome/components/cst328/* @latonita
 esphome/components/cst816/* @clydebarrow
+esphome/components/cst9220/* @clydebarrow
 esphome/components/ct_clamp/* @jesserockz
 esphome/components/current_based/* @djwmarcx
+esphome/components/d01/* @ch604
 esphome/components/dac7678/* @NickB1
 esphome/components/daikin_arc/* @MagicBear
 esphome/components/daikin_brc/* @hagak
@@ -139,12 +157,16 @@ esphome/components/dfplayer/* @glmnet
 esphome/components/dfrobot_sen0395/* @niklasweber
 esphome/components/dht/* @OttoWinter
 esphome/components/display_menu_base/* @numo68
-esphome/components/dlms_meter/* @SimonFischer04
+esphome/components/dlms_meter/* @latonita @PolarGoose @SimonFischer04 @Tomer27cz
 esphome/components/dps310/* @kbx81
 esphome/components/ds1307/* @badbadc0ffee
+esphome/components/ds1603l/* @JakeLC15
 esphome/components/ds2484/* @mrk-its
+esphome/components/ds248x/* @tomwellnitz
+esphome/components/ds3231/* @linkedupbits
 esphome/components/dsmr/* @glmnet @PolarGoose
 esphome/components/duty_time/* @dudanov
+esphome/components/ecocomfort2/* @gledian
 esphome/components/ee895/* @Stock-M
 esphome/components/ektf2232/touchscreen/* @jesserockz
 esphome/components/emc2101/* @ellull
@@ -161,6 +183,7 @@ esphome/components/es7243e/* @kbx81
 esphome/components/es8156/* @kbx81
 esphome/components/es8311/* @kahrendt @kroimon
 esphome/components/es8388/* @P4uLT
+esphome/components/es9038q2m_katana/* @GioMez
 esphome/components/esp32/* @esphome/core
 esphome/components/esp32_ble/* @bdraco @jesserockz @Rapsssito
 esphome/components/esp32_ble_client/* @bdraco @jesserockz
@@ -170,26 +193,34 @@ esphome/components/esp32_camera_web_server/* @ayufan
 esphome/components/esp32_can/* @Sympatron
 esphome/components/esp32_hosted/* @swoboda1337
 esphome/components/esp32_hosted/update/* @swoboda1337
-esphome/components/esp32_improv/* @jesserockz
+esphome/components/esp32_power_management/* @rwrozelle
 esphome/components/esp32_rmt/* @jesserockz
 esphome/components/esp32_rmt_led_strip/* @jesserockz
 esphome/components/esp8266/* @esphome/core
 esphome/components/esp_ldo/* @clydebarrow
+esphome/components/espectre/* @francescopace
+esphome/components/esphome/ota/* @esphome/core
 esphome/components/espnow/* @jesserockz
 esphome/components/espnow/packet_transport/* @EasilyBoredEngineer
 esphome/components/ethernet_info/* @gtjadsonsantos
 esphome/components/event/* @nohat
+esphome/components/exponential_moving_average/* @clydebarrow
 esphome/components/exposure_notifications/* @OttoWinter
 esphome/components/ezo/* @ssieb
 esphome/components/ezo_pmp/* @carlos-sarmiento
+esphome/components/ezsp_proxy_tap/* @kbx81 @puddly
 esphome/components/factory_reset/* @anatoly-savchenkov
 esphome/components/fastled_base/* @OttoWinter
 esphome/components/feedback/* @ianchi
+esphome/components/file/* @esphome/core
+esphome/components/filter_lifetime/* @tuct
 esphome/components/fingerprint_grow/* @alexborro @loongyh @OnFreund
 esphome/components/font/* @clydebarrow @esphome/core
+esphome/components/friedrich/* @rwrozelle
 esphome/components/fs3000/* @kahrendt
 esphome/components/ft5x06/* @clydebarrow
 esphome/components/ft63x6/* @gpambrozio
+esphome/components/fusb302b/* @remcom
 esphome/components/gcja5/* @gcormier
 esphome/components/gdk101/* @Szewcson
 esphome/components/gl_r01_i2c/* @pkejval
@@ -206,13 +237,18 @@ esphome/components/gree/switch/* @nagyrobi
 esphome/components/grove_gas_mc_v2/* @YorkshireIoT
 esphome/components/grove_tb6612fng/* @max246
 esphome/components/growatt_solar/* @leeuwte
+esphome/components/gsl3670/* @clydebarrow
 esphome/components/gt911/* @clydebarrow @jesserockz
 esphome/components/haier/* @paveldn
 esphome/components/haier/binary_sensor/* @paveldn
 esphome/components/haier/button/* @paveldn
+esphome/components/haier/select/* @paveldn
 esphome/components/haier/sensor/* @paveldn
 esphome/components/haier/switch/* @paveldn
 esphome/components/haier/text_sensor/* @paveldn
+esphome/components/haier_tundra/* @fauxpark
+esphome/components/haier_tundra/button/* @fauxpark
+esphome/components/haier_tundra/switch/* @fauxpark
 esphome/components/havells_solar/* @sourabhjaiswal
 esphome/components/hbridge/fan/* @WeekendWarrior
 esphome/components/hbridge/light/* @DotNetDann
@@ -221,6 +257,7 @@ esphome/components/hc8/* @omartijn
 esphome/components/hdc2010/* @optimusprimespace @ssieb
 esphome/components/hdc2080/* @G-Pereira @jesserockz
 esphome/components/hdc302x/* @joshuasing
+esphome/components/hdmi_cec/* @JosVanEijndhoven
 esphome/components/he60r/* @clydebarrow
 esphome/components/heatpumpir/* @rob-deutsch
 esphome/components/hitachi_ac424/* @sourabhjaiswal
@@ -229,9 +266,13 @@ esphome/components/hlw8032/* @rici4kubicek
 esphome/components/hm3301/* @freekode
 esphome/components/hmac_md5/* @dwmw2
 esphome/components/hmac_sha256/* @dwmw2
+esphome/components/hoermann_hcp/* @zweckj
 esphome/components/homeassistant/* @esphome/core @OttoWinter
+esphome/components/homeassistant/button/* @jesserockz
 esphome/components/homeassistant/number/* @landonr
+esphome/components/homeassistant/select/* @jesserockz
 esphome/components/homeassistant/switch/* @Links2004
+esphome/components/homeassistant/text/* @jesserockz
 esphome/components/honeywell_hih_i2c/* @Benichou34
 esphome/components/honeywellabp/* @RubyBailey
 esphome/components/honeywellabp2_i2c/* @jpfaff
@@ -243,6 +284,7 @@ esphome/components/http_request/ota/* @oarcher
 esphome/components/http_request/update/* @jesserockz
 esphome/components/htu31d/* @betterengineering
 esphome/components/hub75/* @stuartparmenter
+esphome/components/husb238/* @latonita
 esphome/components/hydreon_rgxx/* @functionpointer
 esphome/components/hyt271/* @Philippe12
 esphome/components/i2c/* @esphome/core
@@ -250,9 +292,13 @@ esphome/components/i2c_device/* @gabest11
 esphome/components/i2s_audio/* @jesserockz
 esphome/components/i2s_audio/microphone/* @jesserockz
 esphome/components/i2s_audio/speaker/* @jesserockz @kahrendt
+esphome/components/i2s_clockless_led_strip/* @j9brown
 esphome/components/iaqcore/* @yozik04
+esphome/components/icm20689/* @lboue
+esphome/components/icnt86/* @danepowell
 esphome/components/ili9xxx/* @clydebarrow @nielsnl68
 esphome/components/improv_base/* @esphome/core
+esphome/components/improv_ble/* @jesserockz
 esphome/components/improv_serial/* @esphome/core
 esphome/components/ina226/* @latonita @Sergio303
 esphome/components/ina260/* @mreditor97
@@ -265,7 +311,9 @@ esphome/components/inkplate/* @jesserockz @JosipKuci
 esphome/components/integration/* @OttoWinter
 esphome/components/internal_temperature/* @Mat931
 esphome/components/interval/* @esphome/core
+esphome/components/ir_rf_base/* @bdraco @kbx81
 esphome/components/ir_rf_proxy/* @kbx81
+esphome/components/it8951/* @koosoli @limengdu @Passific
 esphome/components/jsn_sr04t/* @Mafus1
 esphome/components/json/* @esphome/core
 esphome/components/kamstrup_kmp/* @cfeenstra1024
@@ -275,22 +323,33 @@ esphome/components/kuntze/* @ssieb
 esphome/components/lc709203f/* @ilikecake
 esphome/components/lcd_menu/* @numo68
 esphome/components/ld2410/* @regevbr @sebcaps
+esphome/components/ld2410s/* @NovakIrs
 esphome/components/ld2412/* @Rihan9
 esphome/components/ld2420/* @descipher
 esphome/components/ld2450/* @hareeshmu
+esphome/components/ld2460/* @jesusvallejo
 esphome/components/ld24xx/* @kbx81
+esphome/components/ld6002b/* @hepter
 esphome/components/ledc/* @OttoWinter
 esphome/components/libretiny/* @kuba2k2
 esphome/components/libretiny_pwm/* @kuba2k2
 esphome/components/light/* @esphome/core
 esphome/components/lightwaverf/* @max246
 esphome/components/lilygo_t5_47/touchscreen/* @jesserockz
+esphome/components/lis2dw12/* @latonita
+esphome/components/lis2dw12_i2c/* @latonita
+esphome/components/lis2dw12_spi/* @latonita
+esphome/components/lis3dh/* @zebble
 esphome/components/lm75b/* @beormund
+esphome/components/ln882h_ble/* @Bl00d-B0b
+esphome/components/ln882h_ble_tracker/* @Bl00d-B0b
 esphome/components/ln882x/* @lamauny
 esphome/components/lock/* @esphome/core
 esphome/components/logger/* @esphome/core
 esphome/components/logger/select/* @clydebarrow
+esphome/components/lp5562/* @DennisGaida
 esphome/components/lps22/* @nagisa
+esphome/components/lsm6ds/* @clydebarrow
 esphome/components/ltr390/* @latonita @sjtrny
 esphome/components/ltr501/* @latonita
 esphome/components/ltr_als_ps/* @latonita
@@ -299,7 +358,9 @@ esphome/components/m5stack_8angle/* @rnauber
 esphome/components/mapping/* @clydebarrow
 esphome/components/matrix_keypad/* @ssieb
 esphome/components/max17043/* @blacknell
+esphome/components/max17048/* @B48D81EFCC
 esphome/components/max31865/* @DAVe3283
+esphome/components/max31888/* @JiriPrchal
 esphome/components/max44009/* @berfenger
 esphome/components/max6956/* @looping40
 esphome/components/max7219digit/* @rspaargaren
@@ -318,7 +379,6 @@ esphome/components/mcp4461/* @p1ngb4ck
 esphome/components/mcp4728/* @berfenger
 esphome/components/mcp47a1/* @jesserockz
 esphome/components/mcp9600/* @mreditor97
-esphome/components/mcp9808/* @k7hpn
 esphome/components/md5/* @esphome/core
 esphome/components/mdns/* @esphome/core
 esphome/components/media_player/* @jesserockz
@@ -329,16 +389,20 @@ esphome/components/microphone/* @jesserockz @kahrendt
 esphome/components/mics_4514/* @jesserockz
 esphome/components/midea/* @dudanov
 esphome/components/midea_ir/* @dudanov
+esphome/components/midea_portasplit/* @Fexiven
 esphome/components/mipi_dsi/* @clydebarrow
 esphome/components/mipi_rgb/* @clydebarrow
 esphome/components/mipi_spi/* @clydebarrow
+esphome/components/mistral_ir/* @lerhum
 esphome/components/mitsubishi/* @RubyBailey
 esphome/components/mitsubishi_cn105/* @crnjan
 esphome/components/mixer/speaker/* @kahrendt
+esphome/components/mk2pvrouter/* @FredM67
 esphome/components/mlx90393/* @functionpointer
 esphome/components/mlx90614/* @jesserockz
 esphome/components/mmc5603/* @benhoff
 esphome/components/mmc5983/* @agoode
+esphome/components/modbus_client/* @exciton
 esphome/components/modbus_controller/* @martgras
 esphome/components/modbus_controller/binary_sensor/* @martgras
 esphome/components/modbus_controller/number/* @martgras
@@ -351,6 +415,7 @@ esphome/components/modbus_server/* @exciton
 esphome/components/mopeka_ble/* @Fabian-Schmidt @spbrogan
 esphome/components/mopeka_pro_check/* @spbrogan
 esphome/components/mopeka_std_check/* @Fabian-Schmidt
+esphome/components/motion/* @esphome/core
 esphome/components/mpl3115a2/* @kbickar
 esphome/components/mpu6886/* @fabaff
 esphome/components/ms8607/* @e28eta
@@ -364,8 +429,11 @@ esphome/components/nextion/switch/* @senexcrenshaw
 esphome/components/nextion/text_sensor/* @senexcrenshaw
 esphome/components/nfc/* @jesserockz @kbx81
 esphome/components/noblex/* @AGalfra
+esphome/components/noise/* @esphome/core
 esphome/components/npi19/* @bakerkj
 esphome/components/nrf52/* @tomaszduda23
+esphome/components/ns2009/* @tchilov
+esphome/components/ns2009/touchscreen/* @tchilov
 esphome/components/number/* @esphome/core
 esphome/components/one_wire/* @ssieb
 esphome/components/online_image/* @clydebarrow @guillempages
@@ -376,12 +444,16 @@ esphome/components/ota/* @esphome/core
 esphome/components/output/* @esphome/core
 esphome/components/packet_transport/* @clydebarrow
 esphome/components/pca6416a/* @Mat931
+esphome/components/pca9505/* @ajfriesen @ssieb
 esphome/components/pca9554/* @bdraco @clydebarrow @hwstar
 esphome/components/pcf85063/* @brogon
 esphome/components/pcf8563/* @KoenBreeman
+esphome/components/pcm5122/* @remcom
+esphome/components/pd_pioneer_ir/* @martinsah
 esphome/components/pi4ioe5v6408/* @jesserockz
 esphome/components/pid/* @OttoWinter
 esphome/components/pipsolar/* @andreashergert1984
+esphome/components/pixoo/* @jesserockz
 esphome/components/pm1006/* @habbie
 esphome/components/pm2005/* @andrewjswan
 esphome/components/pmsa003i/* @sjtrny
@@ -395,12 +467,17 @@ esphome/components/pn7150_i2c/* @jesserockz @kbx81
 esphome/components/pn7160/* @jesserockz @kbx81
 esphome/components/pn7160_i2c/* @jesserockz @kbx81
 esphome/components/pn7160_spi/* @jesserockz @kbx81
+esphome/components/pn71xx/* @jesserockz @kbx81
+esphome/components/power_management/* @rwrozelle
 esphome/components/power_supply/* @esphome/core
 esphome/components/preferences/* @esphome/core
+esphome/components/provisioning/* @esphome/core
 esphome/components/psram/* @esphome/core
 esphome/components/pulse_meter/* @cstaahl @stevebaxter @TrentHouliston
 esphome/components/pvvx_mithermometer/* @pasiz
 esphome/components/pylontech/* @functionpointer
+esphome/components/pzem6l24/* @nuttytree
+esphome/components/qmi8658/* @clydebarrow
 esphome/components/qmp6988/* @andrewpc
 esphome/components/qr_code/* @wjtje
 esphome/components/qspi_dbi/* @clydebarrow
@@ -412,23 +489,28 @@ esphome/components/rc522/* @glmnet
 esphome/components/rc522_i2c/* @glmnet
 esphome/components/rc522_spi/* @glmnet
 esphome/components/rd03d/* @jasstrong
+esphome/components/remko_ar715/* @naKruul
+esphome/components/renogy_inverter_ble/* @emilioaray-dev
 esphome/components/resampler/speaker/* @kahrendt
 esphome/components/restart/* @esphome/core
 esphome/components/rf_bridge/* @jesserockz
 esphome/components/rgbct/* @jesserockz
 esphome/components/ring_buffer/* @kahrendt
 esphome/components/router/speaker/* @kahrendt
-esphome/components/rp2040/* @jesserockz
+esphome/components/rp2/* @jesserockz
 esphome/components/rp2040_ble/* @bdraco
 esphome/components/rp2040_pio_led_strip/* @Papa-DMan
 esphome/components/rp2040_pwm/* @jesserockz
+esphome/components/rp2_ble_tracker/* @bdraco
 esphome/components/rpi_dpi_rgb/* @clydebarrow
 esphome/components/rtl87xx/* @kuba2k2
 esphome/components/rtttl/* @glmnet @ximex
 esphome/components/runtime_image/* @clydebarrow @guillempages @kahrendt
 esphome/components/runtime_stats/* @bdraco
+esphome/components/rx8025t/* @remcom
 esphome/components/rx8130/* @beormund
 esphome/components/safe_mode/* @jsuanet @kbx81 @paulmonigatti
+esphome/components/samsung/* @jorofi
 esphome/components/scd4x/* @martgras @sjtrny
 esphome/components/script/* @esphome/core
 esphome/components/sdl/* @bdm310 @clydebarrow
@@ -442,23 +524,29 @@ esphome/components/select/* @esphome/core
 esphome/components/sen0321/* @notjj
 esphome/components/sen21231/* @shreyaskarnik
 esphome/components/sen5x/* @martgras
-esphome/components/sen6x/* @martgras @mebner86 @mikelawrence @tuct
+esphome/components/sen6x/* @martgras @mebner86 @tuct
 esphome/components/sendspin/* @kahrendt
+esphome/components/sendspin/image/* @kahrendt
 esphome/components/sendspin/media_player/* @kahrendt
 esphome/components/sendspin/media_source/* @kahrendt
 esphome/components/sendspin/sensor/* @kahrendt
+esphome/components/sendspin/switch/* @kahrendt
 esphome/components/sendspin/text_sensor/* @kahrendt
 esphome/components/sensirion_common/* @martgras
 esphome/components/sensor/* @esphome/core
 esphome/components/serial_proxy/* @kbx81
 esphome/components/sfa30/* @ghsensdev
+esphome/components/sfa40/* @NoQuarrel
 esphome/components/sgp40/* @SenexCrenshaw
 esphome/components/sgp4x/* @martgras @SenexCrenshaw
+esphome/components/sh1122_base/* @pl4nkton
+esphome/components/sh1122_spi/* @pl4nkton
 esphome/components/sha256/* @esphome/core
 esphome/components/shelly_dimmer/* @edge90 @rnauber
 esphome/components/sht3xd/* @mrtoy-me
 esphome/components/sht4x/* @sjtrny
 esphome/components/shutdown/* @esphome/core @jsuanet
+esphome/components/siemens_ira211/* @fonix232
 esphome/components/sigma_delta_output/* @Cat-Ion
 esphome/components/sim800l/* @glmnet
 esphome/components/sm10bit_base/* @Cossid
@@ -468,12 +556,14 @@ esphome/components/sm2335/* @Cossid
 esphome/components/sml/* @alengwenus
 esphome/components/smt100/* @piechade
 esphome/components/sn74hc165/* @jesserockz
+esphome/components/snapshot/* @clydebarrow
 esphome/components/socket/* @esphome/core
 esphome/components/sonoff_d1/* @anatoly-savchenkov
 esphome/components/sound_level/* @kahrendt
 esphome/components/spa06_base/* @danielkent-net
 esphome/components/spa06_i2c/* @danielkent-net
 esphome/components/spa06_spi/* @danielkent-net
+esphome/components/spd2010/* @grischard
 esphome/components/speaker/* @jesserockz @kahrendt
 esphome/components/speaker/media_player/* @kahrendt @synesthesiam
 esphome/components/speaker_source/* @kahrendt
@@ -494,6 +584,7 @@ esphome/components/ssd1331_base/* @kbx81
 esphome/components/ssd1331_spi/* @kbx81
 esphome/components/ssd1351_base/* @kbx81
 esphome/components/ssd1351_spi/* @kbx81
+esphome/components/st7123/* @miniskipper
 esphome/components/st7567_base/* @latonita
 esphome/components/st7567_i2c/* @latonita
 esphome/components/st7567_spi/* @latonita
@@ -502,6 +593,7 @@ esphome/components/st7735/* @SenexCrenshaw
 esphome/components/st7789v/* @kbx81
 esphome/components/st7920/* @marsjan155
 esphome/components/statsd/* @Links2004
+esphome/components/stcc4/* @j9brown
 esphome/components/stts22h/* @B48D81EFCC
 esphome/components/substitutions/* @esphome/core
 esphome/components/sun/* @OttoWinter
@@ -512,29 +604,38 @@ esphome/components/sx126x/* @swoboda1337
 esphome/components/sx127x/* @swoboda1337
 esphome/components/sy6970/* @linkedupbits
 esphome/components/syslog/* @clydebarrow
+esphome/components/systa_bus/* @Mat931
 esphome/components/t6615/* @tylermenezes
+esphome/components/tas2780/* @remcom
+esphome/components/tas58xx/* @mrtoy-me @remcom
 esphome/components/tc74/* @sethgirvan
+esphome/components/tca8418/* @zebble
 esphome/components/tca9548a/* @andreashergert1984
 esphome/components/tca9555/* @mobrembski
 esphome/components/tcl112/* @glmnet
+esphome/components/tcp_uart/* @Bascht74
 esphome/components/tee501/* @Stock-M
-esphome/components/teleinfo/* @0hax
+esphome/components/teleinfo/* @0hax @FredM67
 esphome/components/tem3200/* @bakerkj
 esphome/components/template/alarm_control_panel/* @grahambrown11 @hwstar
 esphome/components/template/datetime/* @rfdarter
 esphome/components/template/event/* @nohat
 esphome/components/template/fan/* @ssieb
 esphome/components/text/* @mauritskorse
+esphome/components/tfluna/* @candrews
 esphome/components/thermopro_ble/* @sittner
 esphome/components/thermostat/* @kbx81
 esphome/components/time/* @esphome/core
 esphome/components/tinyusb/* @kbx81
+esphome/components/tinyusb_keyboard/* @frederikbolding
 esphome/components/tlc5947/* @rnauber
 esphome/components/tlc5971/* @IJIJI
 esphome/components/tm1621/* @Philippe12
 esphome/components/tm1637/* @glmnet
 esphome/components/tm1638/* @skykingjwc
 esphome/components/tm1651/* @mrtoy-me
+esphome/components/tmc2209/* @remcom
+esphome/components/tmc22xx/* @remcom
 esphome/components/tmp102/* @timsavage
 esphome/components/tmp1075/* @sybrenstuvel
 esphome/components/tmp117/* @Azimath
@@ -551,14 +652,19 @@ esphome/components/tuya/select/* @bearpawmaxim
 esphome/components/tuya/sensor/* @jesserockz
 esphome/components/tuya/switch/* @jesserockz
 esphome/components/tuya/text_sensor/* @dentra
+esphome/components/tuya/water_heater/* @iago-veiga
 esphome/components/uart/* @esphome/core
 esphome/components/uart/button/* @ssieb
 esphome/components/uart/event/* @eoasmxd
 esphome/components/uart/packet_transport/* @clydebarrow
+esphome/components/uart_mux/* @kbx81
+esphome/components/uart_tcp/* @Bascht74
 esphome/components/udp/* @clydebarrow
 esphome/components/ufire_ec/* @pvizeli
 esphome/components/ufire_ise/* @pvizeli
+esphome/components/ufm01/* @ljungqvist
 esphome/components/ultrasonic/* @ssieb @swoboda1337
+esphome/components/unicode/* @esphome/core
 esphome/components/update/* @jesserockz
 esphome/components/uponor_smatrix/* @kroimon
 esphome/components/usb_cdc_acm/* @kbx81
@@ -574,13 +680,14 @@ esphome/components/wake_on_lan/* @clydebarrow @willwill2will54
 esphome/components/watchdog/* @oarcher
 esphome/components/water_heater/* @dhoeben
 esphome/components/waveshare_epaper/* @clydebarrow
+esphome/components/waveshare_io_ch32v003/* @latonita
 esphome/components/web_server/ota/* @esphome/core
 esphome/components/web_server_base/* @esphome/core
 esphome/components/web_server_idf/* @dentra
 esphome/components/weikai/* @DrCoolZic
 esphome/components/weikai_i2c/* @DrCoolZic
 esphome/components/weikai_spi/* @DrCoolZic
-esphome/components/whirlpool/* @glmnet
+esphome/components/whirlpool/* @arpena @glmnet
 esphome/components/whynter/* @aeonsablaze
 esphome/components/wiegand/* @ssieb
 esphome/components/wireguard/* @droscy @lhoracek @thomas0bernard
@@ -595,10 +702,13 @@ esphome/components/wk2212_spi/* @DrCoolZic
 esphome/components/wl_134/* @hobbypunk90
 esphome/components/wts01/* @alepee
 esphome/components/x9c/* @EtienneMD
+esphome/components/xdb401/* @RT530
 esphome/components/xgzp68xx/* @gcormier
+esphome/components/xiaomi_body_scale/* @dckiller51
 esphome/components/xiaomi_hhccjcy10/* @fariouche
 esphome/components/xiaomi_lywsd02mmc/* @juanluss31
 esphome/components/xiaomi_lywsd03mmc/* @ahpohl
+esphome/components/xiaomi_mccgq02hl/* @ahpohl @morph027
 esphome/components/xiaomi_mhoc303/* @drug123
 esphome/components/xiaomi_mhoc401/* @vevsvevs
 esphome/components/xiaomi_rtcgq02lm/* @jesserockz
@@ -608,6 +718,7 @@ esphome/components/xpt2046/touchscreen/* @nielsnl68 @numo68
 esphome/components/xxtea/* @clydebarrow
 esphome/components/zephyr/* @tomaszduda23
 esphome/components/zephyr_mcumgr/ota/* @tomaszduda23
+esphome/components/zephyr_pwm/* @wiomoc
 esphome/components/zhlt01/* @cfeenstra1024
 esphome/components/zigbee/* @luar123 @tomaszduda23
 esphome/components/zio_ultrasonic/* @kahrendt

Doxyfile

@@ -48,7 +48,7 @@ PROJECT_NAME           = ESPHome
 # could be handy for archiving the generated documentation or if some version
 # control system is used.
 
-PROJECT_NUMBER         = 2026.6.0-dev
+PROJECT_NUMBER         = 2026.11.0-dev
 
 # Using the PROJECT_BRIEF tag one can provide an optional one line description
 # for a project that appears at the top of each page and should give viewer a

MANIFEST.in

@@ -6,3 +6,4 @@ recursive-include esphome *.cpp *.h *.tcc *.c
 recursive-include esphome *.py.script
 recursive-include esphome *.jinja
 recursive-include esphome LICENSE.txt
+recursive-include esphome requirements.txt

THREAT_MODEL.md

@@ -0,0 +1,197 @@
+# ESPHome Threat Model
+
+This document defines the trust boundary for the **ESPHome** repository — the
+Python compiler/CLI and the device firmware it generates — so that real security
+bugs can be told apart from defense-in-depth improvements. It gives contributors,
+reviewers, and security researchers a clear answer to one question:
+**does this issue let an _unauthenticated_ attacker do something they shouldn't?**
+
+Related documents:
+
+- Deployment guidance for operators:
+  https://esphome.io/guides/security_best_practices/
+- The **Device Builder dashboard** (the web UI, its authentication, ingress,
+  Origin/Host gates, and peer-link pairing) lives in a separate repository and
+  has its own threat model. If your report concerns any of that, please read and
+  report there instead:
+  https://github.com/esphome/device-builder/blob/main/docs/THREAT_MODEL.md
+
+## The trust boundary
+
+For this repository there are two trusted inputs by design:
+
+1. **The configuration.** Anyone who can supply or edit a YAML config is trusted
+   (see below).
+2. **Authenticated peers of a running device** — clients holding the device's
+   API/OTA encryption key, API password, OTA password, or web server
+   credentials.
+
+The security boundary is therefore **unauthenticated network traffic vs. those
+trusted inputs.** A bug that lets an unauthenticated attacker cross it is a
+security bug.
+
+## Config authors are host-equivalent by design
+
+Anyone who can supply or edit a configuration is **trusted with full code
+execution on the host that runs `esphome`**, on purpose. This is what the product
+does, not a flaw. A config author can already, through fully supported features:
+
+- Run arbitrary **Python** at validation/compile time via `external_components:`
+  (and other component-import mechanisms) — ESPHome imports those packages as
+  ordinary Python.
+- Run arbitrary **shell** commands through the compile/validate/flash toolchain
+  that ESPHome invokes as subprocesses.
+- Read and write arbitrary files reachable by the process (e.g. via `!include`,
+  `packages:`, `dashboard_import:`, and generated build output).
+
+Because of this, a malicious config author is equivalent to shell access on the
+host running the build.
+
+## What is *not* a security vulnerability
+
+If exploiting an issue requires the ability to supply or edit configuration, it
+is **not** a vulnerability in ESPHome, because that ability already grants host
+code execution. This explicitly includes, among others:
+
+- Template / expression injection in substitutions or any YAML string value
+  (e.g. Jinja `${...}` evaluation reaching Python internals). This grants no
+  capability a config author lacks.
+- `!include` / `packages:` / `dashboard_import:` reading or fetching content
+  from surprising or remote locations.
+- The validator or compiler crashing or behaving unexpectedly on adversarial
+  YAML.
+- ESPHome running as root in the official container — that is the documented
+  deployment posture, reachable by the same caller through the features above.
+
+These do not warrant a CVE or coordinated disclosure. Hardening in these areas
+(for example, sandboxing template evaluation as least-surprise defense-in-depth)
+is welcome as a normal enhancement PR, framed as cleanliness rather than a
+security fix — not as a vulnerability remediation.
+
+## What we do defend
+
+These *are* security bugs in this repo, and we want to hear about them privately:
+
+- Memory-safety or protocol bugs in the generated **device firmware** that are
+  remotely triggerable over the network (native API, web server, OTA, BLE,
+  captive portal, etc.) **without** valid credentials.
+- Authentication or encryption bypass on the device — reaching API calls, OTA
+  updates, or the web server without the configured key/password.
+- Flaws that weaken the device's API or OTA encryption (Noise), OTA auth, or
+  web server auth below their documented guarantees.
+
+## The web server is an open HTTP API by design
+
+The `web_server` component exposes a plain HTTP interface for viewing and
+controlling entities, and, when the `web_server` OTA platform is enabled, for
+uploading firmware at `/update`. Its only access controls are the optional
+`web_server` `auth:` credentials and the network the device sits on.
+
+When `auth:` is not configured, every endpoint is reachable by any client that
+can reach the device. This is intentional; enabling `web_server` without `auth:`
+is choosing an open control surface, in the same way that running native OTA
+without a password leaves OTA open. The API is documented and is meant to be
+called by other devices, scripts, and pages.
+
+As defense-in-depth, the web server checks the `Origin` header on browser requests
+to its entity control and state endpoints: a request whose `Origin` does not match
+the address the device is served on is rejected, and the `allowed_origins` option
+widens that list. This blocks the common "confused deputy" (CSRF) case where a page
+the operator visits drives the device through their browser. It is **not** an
+authentication boundary: it only constrains browsers. Any client that omits the
+`Origin` header — `curl`, scripts, or other non-browser callers on the same
+network — reaches every endpoint exactly as before. The check also does not cover
+the web OTA `/update` endpoint. The device performs no CSRF-token or `Referer`
+validation. The following are therefore **not** vulnerabilities in this repository:
+
+- Requests without an `Origin` header (for example `curl`) reaching the control
+  endpoints, whether or not `web_server` `auth:` is set.
+- Requests from an origin the operator added to `allowed_origins`.
+- Cross-origin or CSRF firmware upload through the web OTA endpoint (`/update`) when
+  web OTA is enabled without `web_server` `auth:`. The `/update` endpoint is not
+  covered by the `Origin` check; this is the same exposure as running OTA without a
+  password.
+
+The supported defenses are `web_server` `auth:`, protecting OTA (a web password or
+a native OTA password), and keeping devices on a trusted, segmented network. See
+the security best practices guide linked above.
+
+What remains in scope is bypassing `web_server` `auth:` when it *is* configured,
+and any memory-safety or protocol bug in the server reachable without credentials.
+
+This section documents the current design and scope; it is not a judgment that the
+design is optimal or that it will not change.
+
+## OTA update encryption
+
+The `esphome` OTA platform optionally encrypts updates with the same Noise
+`NNpsk0` pattern the native API uses; one key protects the device. A device
+whose `api:` block has an encryption key, static in the YAML or provisioned at
+runtime, compiles in the transport and offers it on every OTA connection once
+it holds a key, so an uploader presenting that key gets the guarantees below
+even without an `ota: encryption:` block; only that block makes the device
+require encryption. The guarantees are: the firmware image is confidential in
+transit, the uploader is authenticated by the pre-shared key, and the plaintext
+negotiation preceding the handshake is bound into the handshake prologue, so
+stripping or tampering with it fails the first MAC. With `ota: encryption:`
+configured both ends fail closed with no override: the device refuses
+plaintext uploads, and the CLI refuses to send plaintext when a key is
+configured. Without that block the CLI tries a static api key when the device
+offers and, until 2027.3.0, falls back to plaintext with a warning when the
+offer is missing or the handshake fails; a runtime provisioned key never
+reaches the CLI, so those uploads stay plaintext.
+
+Defeating any of that without the key is in scope: a device that requires
+encryption accepting a plaintext or downgraded upload, getting past the MAC,
+or recovering image contents from captured traffic.
+
+The following are **not** vulnerabilities, by design:
+
+- Plaintext OTA on a device with no `ota: encryption:` block, including one
+  that offers encryption because it has an api key. That is the documented
+  default, authenticated (if at all) by the OTA password. An uploader that
+  takes the offer skips the password; the key authenticates it. With a
+  runtime provisioned key and no `provisioning:` window, whoever provisions
+  the key gains that upload path too; validation warns about the pair.
+- The CLI plaintext fallback until 2027.3.0: without `ota: encryption:` an
+  active attacker who strips the offer or breaks the handshake can make a
+  keyed CLI upload plaintext, with the pre-existing plaintext exposure. A
+  device that requires encryption still refuses that upload.
+- The enablement window: firmware built with a static api key already offers
+  encryption, so turning on `ota: encryption:` is itself an encrypted upload.
+  Older firmware needs one last plaintext upload of an offering build, with
+  the pre-existing plaintext exposure.
+- The web OTA `/update` endpoint alongside encryption. With the `web_server`
+  or `prometheus` component the shared listener is always up, so the endpoint
+  stays reachable and validation warns about that combination;
+  `captive_portal:` alone brings the listener up only for the fallback AP
+  window, which is the intended recovery path, so that is not warned about.
+- CLI retry behavior on transport or MAC failures; every attempt renegotiates
+  a fresh handshake with fresh ephemerals, so retrying does not weaken
+  authentication.
+
+## Explicitly out of scope
+
+- Local attackers who already have shell access on the host that runs `esphome`.
+- Supply-chain attacks against ESPHome or its dependencies.
+- Operator-supplied hostile YAML (covered above — config authoring is trusted).
+- Attacks that require an already-authenticated device peer (someone who already
+  holds the API key / OTA / web credentials).
+- Access to the device web server or its web OTA endpoint by non-browser clients
+  (those that send no `Origin` header). The web server is an open HTTP API by
+  design (see above); browser cross-origin requests are blocked by default, but the
+  real controls are `web_server` `auth:` and network isolation.
+- Anything in the dashboard / device-builder — report that in its own repository
+  (linked at the top).
+- Deployments where the operator removed protections or exposed credentials. See
+  the security best practices guide:
+  https://esphome.io/guides/security_best_practices/
+
+## Reporting a vulnerability
+
+If you believe you've found an issue that crosses the unauthenticated boundary
+above, please report it privately via GitHub Security Advisories rather than a
+public issue. For issues that require config-write access, please review this
+document first — they are very likely out of scope by design. For dashboard /
+device-builder issues, report against that repository and consult its threat
+model (linked at the top).

codecov.yml

@@ -0,0 +1,18 @@
+coverage:
+  status:
+    patch:
+      default:
+        target: 100%
+        threshold: 0%
+    project:
+      default:
+        informational: true
+
+ignore:
+  - "esphome/components/**/*"
+  - "esphome/analyze_memory/**/*"
+  - "tests/integration/**/*"
+
+comment:
+  layout: "reach, diff, flags, files"
+  require_changes: true

docker/Dockerfile

@@ -1,10 +1,9 @@
 ARG BUILD_VERSION=dev
-ARG BUILD_OS=alpine
-ARG BUILD_BASE_VERSION=2025.04.0
+ARG BUILD_BASE_VERSION=2026.06.1
 ARG BUILD_TYPE=docker
 
-FROM ghcr.io/esphome/docker-base:${BUILD_OS}-${BUILD_BASE_VERSION} AS base-source-docker
-FROM ghcr.io/esphome/docker-base:${BUILD_OS}-ha-addon-${BUILD_BASE_VERSION} AS base-source-ha-addon
+FROM ghcr.io/esphome/docker-base:debian-${BUILD_BASE_VERSION} AS base-source-docker
+FROM ghcr.io/esphome/docker-base:debian-ha-addon-${BUILD_BASE_VERSION} AS base-source-ha-addon
 
 ARG BUILD_TYPE
 FROM base-source-${BUILD_TYPE} AS base
@@ -12,20 +11,6 @@ FROM base-source-${BUILD_TYPE} AS base
 RUN git config --system --add safe.directory "*" \
     && git config --system advice.detachedHead false
 
-# Install build tools for Python packages that require compilation
-# (e.g., ruamel.yaml.clib used by ESP-IDF's idf-component-manager).
-# Also install libusb-1.0 at runtime so the ESP-IDF tools installer can
-# validate openocd-esp32 (it dynamically links libusb-1.0.so.0); without
-# it idf_tools.py rejects the openocd install with exit 127 and aborts
-# the whole framework setup.
-RUN if command -v apk > /dev/null; then \
-        apk add --no-cache build-base libusb; \
-    else \
-        apt-get update \
-        && apt-get install -y --no-install-recommends build-essential libusb-1.0-0 \
-        && rm -rf /var/lib/apt/lists/*; \
-    fi
-
 ENV PIP_DISABLE_PIP_VERSION_CHECK=1
 
 RUN pip install --no-cache-dir -U pip uv==0.10.1
@@ -36,14 +21,14 @@ RUN \
     uv pip install --no-cache-dir \
     -r /requirements.txt
 
+# Install the ESPHome Device Builder dashboard.
+RUN uv pip install --no-cache-dir esphome-device-builder==1.23.1
+
 RUN \
     platformio settings set enable_telemetry No \
     && platformio settings set check_platformio_interval 1000000 \
     && mkdir -p /piolibs
 
-COPY script/platformio_install_deps.py platformio.ini /
-RUN /platformio_install_deps.py /platformio.ini --libraries
-
 ARG BUILD_VERSION
 
 LABEL \

docker/build.py

@@ -20,6 +20,10 @@ TYPE_HA_ADDON = "ha-addon"
 TYPE_LINT = "lint"
 TYPES = [TYPE_DOCKER, TYPE_HA_ADDON, TYPE_LINT]
 
+REGISTRY_GHCR = "ghcr"
+REGISTRY_DOCKERHUB = "dockerhub"
+REGISTRIES = [REGISTRY_GHCR, REGISTRY_DOCKERHUB]
+
 
 parser = argparse.ArgumentParser()
 parser.add_argument(
@@ -34,6 +38,12 @@ parser.add_argument(
 parser.add_argument(
     "--build-type", choices=TYPES, required=True, help="The type of build to run"
 )
+parser.add_argument(
+    "--registry",
+    choices=REGISTRIES,
+    action="append",
+    help="Restrict to specific registries (default: all). May be passed multiple times.",
+)
 parser.add_argument(
     "--dry-run", action="store_true", help="Don't run any commands, just print them"
 )
@@ -45,6 +55,11 @@ build_parser.add_argument("--push", help="Also push the images", action="store_t
 build_parser.add_argument(
     "--load", help="Load the docker image locally", action="store_true"
 )
+build_parser.add_argument(
+    "--no-cache-to",
+    help="Don't write the build cache (avoids polluting the shared cache)",
+    action="store_true",
+)
 manifest_parser = subparsers.add_parser(
     "manifest", help="Create a manifest from already pushed images"
 )
@@ -95,11 +110,14 @@ def main():
                 print("Command failed")
                 sys.exit(1)
 
+    registries = args.registry or REGISTRIES
+
     # detect channel from tag
     match = re.match(r"^(\d+\.\d+)(?:\.\d+)?(b\d+)?$", args.tag)
     major_minor_version = None
     if match is None:
-        channel = CHANNEL_DEV
+        # Custom tag (e.g. a branch name) -- push only the tag itself
+        channel = None
     elif match.group(2) is None:
         major_minor_version = match.group(1)
         channel = CHANNEL_RELEASE
@@ -128,11 +146,18 @@ def main():
             CHANNEL_DEV: "cache-dev",
             CHANNEL_BETA: "cache-beta",
             CHANNEL_RELEASE: "cache-latest",
-        }[channel]
-        cache_img = f"ghcr.io/{params.build_to}:{cache_tag}"
+        }.get(channel, "cache-dev")
+        # Cache images live alongside the pushed images; prefer GHCR when it is
+        # one of the selected registries, otherwise fall back to Docker Hub so a
+        # registry-restricted build doesn't need GHCR auth.
+        cache_prefix = "ghcr.io/" if REGISTRY_GHCR in registries else ""
+        cache_img = f"{cache_prefix}{params.build_to}:{cache_tag}"
 
-        imgs = [f"{params.build_to}:{tag}" for tag in tags_to_push]
-        imgs += [f"ghcr.io/{params.build_to}:{tag}" for tag in tags_to_push]
+        imgs = []
+        if REGISTRY_DOCKERHUB in registries:
+            imgs += [f"{params.build_to}:{tag}" for tag in tags_to_push]
+        if REGISTRY_GHCR in registries:
+            imgs += [f"ghcr.io/{params.build_to}:{tag}" for tag in tags_to_push]
 
         # 3. build
         cmd = [
@@ -155,7 +180,9 @@ def main():
         for img in imgs:
             cmd += ["--tag", img]
         if args.push:
-            cmd += ["--push", "--cache-to", f"type=registry,ref={cache_img},mode=max"]
+            cmd += ["--push"]
+            if not args.no_cache_to:
+                cmd += ["--cache-to", f"type=registry,ref={cache_img},mode=max"]
         if args.load:
             cmd += ["--load"]
 
@@ -163,20 +190,22 @@ def main():
     elif args.command == "manifest":
         manifest = DockerParams.for_type_arch(args.build_type, ARCH_AMD64).manifest_to
 
-        targets = [f"{manifest}:{tag}" for tag in tags_to_push]
-        targets += [f"ghcr.io/{manifest}:{tag}" for tag in tags_to_push]
-        # 1. Create manifests
+        targets = []
+        if REGISTRY_DOCKERHUB in registries:
+            targets += [f"{manifest}:{tag}" for tag in tags_to_push]
+        if REGISTRY_GHCR in registries:
+            targets += [f"ghcr.io/{manifest}:{tag}" for tag in tags_to_push]
+        # Use buildx imagetools (not `docker manifest`) so the per-arch sources,
+        # which buildx pushes as single-platform manifest lists, are combined
+        # and pushed correctly in one step.
         for target in targets:
-            cmd = ["docker", "manifest", "create", target]
+            cmd = ["docker", "buildx", "imagetools", "create", "--tag", target]
             for arch in ARCHS:
                 src = f"{DockerParams.for_type_arch(args.build_type, arch).build_to}:{args.tag}"
                 if target.startswith("ghcr.io"):
                     src = f"ghcr.io/{src}"
                 cmd.append(src)
             run_command(*cmd)
-        # 2. Push manifests
-        for target in targets:
-            run_command("docker", "manifest", "push", target)
 
 
 if __name__ == "__main__":

docker/docker_entrypoint.sh

@@ -21,10 +21,27 @@ export PLATFORMIO_PLATFORMS_DIR="${pio_cache_base}/platforms"
 export PLATFORMIO_PACKAGES_DIR="${pio_cache_base}/packages"
 export PLATFORMIO_CACHE_DIR="${pio_cache_base}/cache"
 
+# Keep the native toolchain installs and compiler caches on the persistent
+# cache root, not the container's user cache dir: it is lost on every
+# restart, and not writable when the container runs as a non-root user.
+export ESPHOME_ESP_IDF_PREFIX="$(dirname "${pio_cache_base}")/idf"
+export ESPHOME_SDK_NRF_PREFIX="$(dirname "${pio_cache_base}")/sdk-nrf"
+export ESPHOME_ARDUINO8266_PREFIX="$(dirname "${pio_cache_base}")/arduino8266"
+export ESPHOME_HOST_PREFIX="$(dirname "${pio_cache_base}")/host"
+export ESPHOME_PLATFORMIO_CCACHE_DIR="$(dirname "${pio_cache_base}")/platformio-ccache"
+
 # If /build is mounted, use that as the build path
 # otherwise use path in /config (so that builds aren't lost on container restart)
 if [[ -d /build ]]; then
     export ESPHOME_BUILD_PATH=/build
 fi
 
+# The default CMD is "dashboard /config". Route the dashboard to the new
+# Device Builder, but pass every other subcommand (compile, run, config,
+# logs, ...) straight through to the esphome CLI so direct CLI use keeps working.
+if [[ "$1" == "dashboard" ]]; then
+    shift
+    exec esphome-device-builder "$@"
+fi
+
 exec esphome "$@"

docker/generate_tags.py

@@ -1,5 +1,6 @@
 #!/usr/bin/env python3
 import argparse
+import os
 import re
 
 CHANNEL_DEV = "dev"
@@ -64,7 +65,10 @@ def main():
 
     suffix = f"-{args.suffix}" if args.suffix else ""
 
-    image_name = f"esphome/esphome{suffix}"
+    repository = (
+        (os.environ.get("GITHUB_REPOSITORY") or "esphome/esphome").strip().lower()
+    )
+    image_name = f"{repository}{suffix}"
 
     print(f"channel={channel}")
 

docker/ha-addon-rootfs/etc/cont-init.d/40-device-builder.sh

@@ -1,22 +0,0 @@
-#!/usr/bin/with-contenv bashio
-# ==============================================================================
-# Installs the latest prerelease of esphome-device-builder when the
-# `use_new_device_builder` config option is enabled.
-# This is a temporary install-on-boot step until esphome-device-builder
-# becomes a direct dependency of esphome.
-# ==============================================================================
-
-if ! bashio::config.true 'use_new_device_builder'; then
-  exit 0
-fi
-
-bashio::log.info "Installing latest prerelease of esphome-device-builder..."
-if command -v uv > /dev/null; then
-  uv pip install --system --no-cache-dir --prerelease=allow --upgrade \
-    esphome-device-builder ||
-    bashio::exit.nok "Failed installing esphome-device-builder."
-else
-  pip install --no-cache-dir --pre --upgrade esphome-device-builder ||
-    bashio::exit.nok "Failed installing esphome-device-builder."
-fi
-bashio::log.info "Installed esphome-device-builder."

docker/ha-addon-rootfs/etc/nginx/includes/mime.types

@@ -1,96 +0,0 @@
-types {
-    text/html                                        html htm shtml;
-    text/css                                         css;
-    text/xml                                         xml;
-    image/gif                                        gif;
-    image/jpeg                                       jpeg jpg;
-    application/javascript                           js;
-    application/atom+xml                             atom;
-    application/rss+xml                              rss;
-
-    text/mathml                                      mml;
-    text/plain                                       txt;
-    text/vnd.sun.j2me.app-descriptor                 jad;
-    text/vnd.wap.wml                                 wml;
-    text/x-component                                 htc;
-
-    image/png                                        png;
-    image/svg+xml                                    svg svgz;
-    image/tiff                                       tif tiff;
-    image/vnd.wap.wbmp                               wbmp;
-    image/webp                                       webp;
-    image/x-icon                                     ico;
-    image/x-jng                                      jng;
-    image/x-ms-bmp                                   bmp;
-
-    font/woff                                        woff;
-    font/woff2                                       woff2;
-
-    application/java-archive                         jar war ear;
-    application/json                                 json;
-    application/mac-binhex40                         hqx;
-    application/msword                               doc;
-    application/pdf                                  pdf;
-    application/postscript                           ps eps ai;
-    application/rtf                                  rtf;
-    application/vnd.apple.mpegurl                    m3u8;
-    application/vnd.google-earth.kml+xml             kml;
-    application/vnd.google-earth.kmz                 kmz;
-    application/vnd.ms-excel                         xls;
-    application/vnd.ms-fontobject                    eot;
-    application/vnd.ms-powerpoint                    ppt;
-    application/vnd.oasis.opendocument.graphics      odg;
-    application/vnd.oasis.opendocument.presentation  odp;
-    application/vnd.oasis.opendocument.spreadsheet   ods;
-    application/vnd.oasis.opendocument.text          odt;
-    application/vnd.openxmlformats-officedocument.presentationml.presentation
-                                                     pptx;
-    application/vnd.openxmlformats-officedocument.spreadsheetml.sheet
-                                                     xlsx;
-    application/vnd.openxmlformats-officedocument.wordprocessingml.document
-                                                     docx;
-    application/vnd.wap.wmlc                         wmlc;
-    application/x-7z-compressed                      7z;
-    application/x-cocoa                              cco;
-    application/x-java-archive-diff                  jardiff;
-    application/x-java-jnlp-file                     jnlp;
-    application/x-makeself                           run;
-    application/x-perl                               pl pm;
-    application/x-pilot                              prc pdb;
-    application/x-rar-compressed                     rar;
-    application/x-redhat-package-manager             rpm;
-    application/x-sea                                sea;
-    application/x-shockwave-flash                    swf;
-    application/x-stuffit                            sit;
-    application/x-tcl                                tcl tk;
-    application/x-x509-ca-cert                       der pem crt;
-    application/x-xpinstall                          xpi;
-    application/xhtml+xml                            xhtml;
-    application/xspf+xml                             xspf;
-    application/zip                                  zip;
-
-    application/octet-stream                         bin exe dll;
-    application/octet-stream                         deb;
-    application/octet-stream                         dmg;
-    application/octet-stream                         iso img;
-    application/octet-stream                         msi msp msm;
-
-    audio/midi                                       mid midi kar;
-    audio/mpeg                                       mp3;
-    audio/ogg                                        ogg;
-    audio/x-m4a                                      m4a;
-    audio/x-realaudio                                ra;
-
-    video/3gpp                                       3gpp 3gp;
-    video/mp2t                                       ts;
-    video/mp4                                        mp4;
-    video/mpeg                                       mpeg mpg;
-    video/quicktime                                  mov;
-    video/webm                                       webm;
-    video/x-flv                                      flv;
-    video/x-m4v                                      m4v;
-    video/x-mng                                      mng;
-    video/x-ms-asf                                   asx asf;
-    video/x-ms-wmv                                   wmv;
-    video/x-msvideo                                  avi;
-}

docker/ha-addon-rootfs/etc/nginx/includes/proxy_params.conf

@@ -1,16 +0,0 @@
-proxy_http_version          1.1;
-proxy_ignore_client_abort   off;
-proxy_read_timeout          86400s;
-proxy_redirect              off;
-proxy_send_timeout          86400s;
-proxy_max_temp_file_size    0;
-
-proxy_set_header Accept-Encoding "";
-proxy_set_header Connection $connection_upgrade;
-proxy_set_header Host $http_host;
-proxy_set_header Upgrade $http_upgrade;
-proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
-proxy_set_header X-Forwarded-Proto $scheme;
-proxy_set_header X-NginX-Proxy true;
-proxy_set_header X-Real-IP $remote_addr;
-proxy_set_header Authorization "";

docker/ha-addon-rootfs/etc/nginx/includes/server_params.conf

@@ -1,8 +0,0 @@
-root            /dev/null;
-server_name     $hostname;
-
-client_max_body_size 512m;
-
-add_header X-Content-Type-Options nosniff;
-add_header X-XSS-Protection "1; mode=block";
-add_header X-Robots-Tag none;

docker/ha-addon-rootfs/etc/nginx/includes/ssl_params.conf

@@ -1,8 +0,0 @@
-ssl_protocols TLSv1.2 TLSv1.3;
-ssl_prefer_server_ciphers off;
-ssl_ciphers ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:DHE-RSA-AES128-GCM-SHA256:DHE-RSA-AES256-GCM-SHA384;
-ssl_session_timeout  10m;
-ssl_session_cache shared:SSL:10m;
-ssl_session_tickets off;
-ssl_stapling on;
-ssl_stapling_verify on;

docker/ha-addon-rootfs/etc/nginx/includes/upstream.conf

@@ -1,3 +0,0 @@
-upstream esphome {
-    server unix:/var/run/esphome.sock;
-}

docker/ha-addon-rootfs/etc/nginx/nginx.conf

@@ -1,30 +0,0 @@
-daemon off;
-user root;
-pid /var/run/nginx.pid;
-worker_processes 1;
-error_log /proc/1/fd/1 error;
-events {
-    worker_connections 1024;
-}
-
-http {
-    include /etc/nginx/includes/mime.types;
-
-    access_log              off;
-    default_type            application/octet-stream;
-    gzip                    on;
-    keepalive_timeout       65;
-    sendfile                on;
-    server_tokens           off;
-
-    tcp_nodelay             on;
-    tcp_nopush              on;
-
-    map $http_upgrade $connection_upgrade {
-        default upgrade;
-        ''      close;
-    }
-
-    include /etc/nginx/includes/upstream.conf;
-    include /etc/nginx/servers/*.conf;
-}

docker/ha-addon-rootfs/etc/nginx/servers/.gitkeep

@@ -1 +0,0 @@
-Without requirements or design, programming is the art of adding bugs to an empty text file. (Louis Srygley)

docker/ha-addon-rootfs/etc/nginx/templates/direct.gtpl

@@ -1,28 +0,0 @@
-server {
-    {{ if not .ssl }}
-    listen 6052 default_server;
-    {{ else }}
-    listen 6052 default_server ssl http2;
-    {{ end }}
-
-    include /etc/nginx/includes/server_params.conf;
-    include /etc/nginx/includes/proxy_params.conf;
-
-    {{ if .ssl }}
-    include /etc/nginx/includes/ssl_params.conf;
-
-    ssl_certificate /ssl/{{ .certfile }};
-    ssl_certificate_key /ssl/{{ .keyfile }};
-
-    # Redirect http requests to https on the same port.
-    # https://rageagainstshell.com/2016/11/redirect-http-to-https-on-the-same-port-in-nginx/
-    error_page 497 https://$http_host$request_uri;
-    {{ end }}
-
-    # Clear Home Assistant Ingress header
-    proxy_set_header X-HA-Ingress "";
-
-    location / {
-        proxy_pass http://esphome;
-    }
-}

docker/ha-addon-rootfs/etc/nginx/templates/ingress.gtpl

@@ -1,18 +0,0 @@
-server {
-    listen 127.0.0.1:{{ .port }} default_server;
-    listen {{ .interface }}:{{ .port }} default_server;
-
-    include /etc/nginx/includes/server_params.conf;
-    include /etc/nginx/includes/proxy_params.conf;
-
-    # Set Home Assistant Ingress header
-    proxy_set_header X-HA-Ingress "YES";
-
-    location / {
-        allow   172.30.32.2;
-        allow   127.0.0.1;
-        deny    all;
-
-        proxy_pass http://esphome;
-    }
-}

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/discovery/run

@@ -4,7 +4,6 @@
 # Home Assistant Add-on: ESPHome
 # Sends discovery information to Home Assistant.
 # ==============================================================================
-declare config
 declare port
 
 # We only disable it when disabled explicitly
@@ -16,17 +15,20 @@ fi
 
 port=$(bashio::addon.ingress_port)
 
-# Wait for NGINX to become available
+# Wait for the ESPHome Device Builder to become available
 bashio::net.wait_for "${port}" "127.0.0.1" 300
 
-config=$(\
-    bashio::var.json \
-        host "127.0.0.1" \
-        port "^${port}" \
-)
+# Send one discovery message; the config is a JSON string built with bashio::var.json.
+send_discovery() {
+    local service=$1
+    local config=$2
+    if bashio::discovery "${service}" "${config}" > /dev/null; then
+        bashio::log.info "Successfully send ${service} discovery information to Home Assistant."
+    else
+        bashio::log.error "${service} discovery message to Home Assistant failed!"
+    fi
+}
 
-if bashio::discovery "esphome" "${config}" > /dev/null; then
-    bashio::log.info "Successfully send discovery information to Home Assistant."
-else
-    bashio::log.error "Discovery message to Home Assistant failed!"
-fi
+send_discovery "esphome" "$(bashio::var.json host "127.0.0.1" port "^${port}")"
+# The Device Builder MCP server, consumed by Home Assistant's mcp integration.
+send_discovery "mcp" "$(bashio::var.json url "http://127.0.0.1:${port}/api/mcp")"

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/esphome/finish

@@ -2,7 +2,7 @@
 # shellcheck shell=bash
 # ==============================================================================
 # Home Assistant Community Add-on: ESPHome
-# Take down the S6 supervision tree when ESPHome dashboard fails
+# Take down the S6 supervision tree when ESPHome Device Builder fails
 # ==============================================================================
 declare exit_code
 readonly exit_code_container=$(</run/s6-linux-init-container-results/exitcode)
@@ -10,7 +10,7 @@ readonly exit_code_service="${1}"
 readonly exit_code_signal="${2}"
 
 bashio::log.info \
-  "Service ESPHome dashboard exited with code ${exit_code_service}" \
+  "Service ESPHome Device Builder exited with code ${exit_code_service}" \
   "(by signal ${exit_code_signal})"
 
 if [[ "${exit_code_service}" -eq 256 ]]; then

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/esphome/run

@@ -2,7 +2,7 @@
 # shellcheck shell=bash
 # ==============================================================================
 # Community Hass.io Add-ons: ESPHome
-# Runs the ESPHome dashboard
+# Runs the ESPHome Device Builder
 # ==============================================================================
 readonly pio_cache_base=/data/cache/platformio
 
@@ -15,18 +15,19 @@ export PLATFORMIO_PLATFORMS_DIR="${pio_cache_base}/platforms"
 export PLATFORMIO_PACKAGES_DIR="${pio_cache_base}/packages"
 export PLATFORMIO_CACHE_DIR="${pio_cache_base}/cache"
 
+# Keep the native toolchain installs and compiler caches on the persistent
+# /data volume, not the container's ephemeral user cache dir (wiped on every
+# add-on update/restart).
+export ESPHOME_ESP_IDF_PREFIX=/data/cache/idf
+export ESPHOME_SDK_NRF_PREFIX=/data/cache/sdk-nrf
+export ESPHOME_ARDUINO8266_PREFIX=/data/cache/arduino8266
+export ESPHOME_HOST_PREFIX=/data/cache/host
+export ESPHOME_PLATFORMIO_CCACHE_DIR=/data/cache/platformio-ccache
+
 if bashio::config.true 'leave_front_door_open'; then
     export DISABLE_HA_AUTHENTICATION=true
 fi
 
-if bashio::config.true 'streamer_mode'; then
-    export ESPHOME_STREAMER_MODE=true
-fi
-
-if bashio::config.has_value 'relative_url'; then
-    export ESPHOME_DASHBOARD_RELATIVE_URL=$(bashio::config 'relative_url')
-fi
-
 if bashio::config.has_value 'default_compile_process_limit'; then
     export ESPHOME_DEFAULT_COMPILE_PROCESS_LIMIT=$(bashio::config 'default_compile_process_limit')
 else
@@ -49,12 +50,21 @@ if bashio::fs.directory_exists '/config/esphome/.esphome'; then
     rm -rf /config/esphome/.esphome
 fi
 
-if bashio::config.true 'use_new_device_builder'; then
-    bashio::log.info "Starting ESPHome Device Builder..."
-    exec esphome-device-builder /config/esphome \
-        --ha-addon \
-        --ingress-port "$(bashio::addon.ingress_port)"
+# Only signal device-builder to expose the public LAN port when the operator
+# mapped port 6052, matching the legacy dashboard where nginx listened on the
+# fixed port 6052 only when it was configured. We use the mapping purely as a
+# presence check and don't forward the published value; device-builder binds
+# its default port 6052 (the fixed container port, as the legacy
+# "listen 6052" did). --ha-addon-allow-public is inert on its own: the no-auth
+# gate is the DISABLE_HA_AUTHENTICATION env var set above, so both opt-ins are
+# required to bind 6052 unauthenticated; either alone stays ingress-only.
+set --
+if bashio::var.has_value "$(bashio::addon.port 6052)"; then
+    set -- --ha-addon-allow-public
 fi
 
-bashio::log.info "Starting ESPHome dashboard..."
-exec esphome dashboard /config/esphome --socket /var/run/esphome.sock --ha-addon
+bashio::log.info "Starting ESPHome Device Builder..."
+exec esphome-device-builder /config/esphome \
+    --ha-addon \
+    --ingress-port "$(bashio::addon.ingress_port)" \
+    "$@"

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/init-nginx/run

@@ -1,35 +0,0 @@
-#!/command/with-contenv bashio
-# shellcheck shell=bash
-# ==============================================================================
-# Community Hass.io Add-ons: ESPHome
-# Configures NGINX for use with ESPHome
-# ==============================================================================
-
-# When the new device builder is enabled it serves HA ingress directly,
-# so nginx is not used at all -- skip configuration.
-if bashio::config.true 'use_new_device_builder'; then
-    bashio::log.info "Skipping NGINX setup: new device builder serves ingress directly."
-    bashio::exit.ok
-fi
-
-mkdir -p /var/log/nginx
-
-# Generate Ingress configuration
-bashio::var.json \
-    interface "$(bashio::addon.ip_address)" \
-    port "^$(bashio::addon.ingress_port)" \
-    | tempio \
-        -template /etc/nginx/templates/ingress.gtpl \
-        -out /etc/nginx/servers/ingress.conf
-
-# Generate direct access configuration, if enabled.
-if bashio::var.has_value "$(bashio::addon.port 6052)"; then
-    bashio::config.require.ssl
-    bashio::var.json \
-        certfile "$(bashio::config 'certfile')" \
-        keyfile "$(bashio::config 'keyfile')" \
-        ssl "^$(bashio::config 'ssl')" \
-        | tempio \
-            -template /etc/nginx/templates/direct.gtpl \
-            -out /etc/nginx/servers/direct.conf
-fi

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/init-nginx/type

@@ -1 +0,0 @@
-oneshot

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/init-nginx/up

@@ -1 +0,0 @@
-/etc/s6-overlay/s6-rc.d/init-nginx/run

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/nginx/finish

@@ -1,25 +0,0 @@
-#!/command/with-contenv bashio
-# ==============================================================================
-# Community Hass.io Add-ons: ESPHome
-# Take down the S6 supervision tree when NGINX fails
-# ==============================================================================
-declare exit_code
-readonly exit_code_container=$(</run/s6-linux-init-container-results/exitcode)
-readonly exit_code_service="${1}"
-readonly exit_code_signal="${2}"
-
-bashio::log.info \
-  "Service NGINX exited with code ${exit_code_service}" \
-  "(by signal ${exit_code_signal})"
-
-if [[ "${exit_code_service}" -eq 256 ]]; then
-  if [[ "${exit_code_container}" -eq 0 ]]; then
-    echo $((128 + $exit_code_signal)) > /run/s6-linux-init-container-results/exitcode
-  fi
-  [[ "${exit_code_signal}" -eq 15 ]] && exec /run/s6/basedir/bin/halt
-elif [[ "${exit_code_service}" -ne 0 ]]; then
-  if [[ "${exit_code_container}" -eq 0 ]]; then
-    echo "${exit_code_service}" > /run/s6-linux-init-container-results/exitcode
-  fi
-  exec /run/s6/basedir/bin/halt
-fi

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/nginx/run

@@ -1,23 +0,0 @@
-#!/command/with-contenv bashio
-# shellcheck shell=bash
-# ==============================================================================
-# Community Hass.io Add-ons: ESPHome
-# Runs the NGINX proxy
-# ==============================================================================
-
-# The new device builder handles HA ingress itself, so nginx is bypassed.
-# Block the longrun forever so s6 keeps the dependency satisfied and does
-# not respawn it.
-if bashio::config.true 'use_new_device_builder'; then
-    bashio::log.info "NGINX bypassed: new device builder serves ingress directly."
-    exec sleep infinity
-fi
-
-bashio::log.info "Waiting for ESPHome dashboard to come up..."
-
-while [[ ! -S /var/run/esphome.sock ]]; do
-  sleep 0.5
-done
-
-bashio::log.info "Starting NGINX..."
-exec nginx

docker/ha-addon-rootfs/etc/s6-overlay/s6-rc.d/nginx/type

@@ -1 +0,0 @@
-longrun

docker/test_configs/bk72xx-arduino.yaml

@@ -0,0 +1,19 @@
+esphome:
+  name: docker-test-bk72xx-arduino
+
+bk72xx:
+  board: generic-bk7231n-qfn32-tuya
+
+logger:
+
+wifi:
+  ssid: MySSID
+  password: password1
+  ap:
+
+# mqtt and captive_portal together pull in AsyncTCP and ESPAsyncWebServer;
+# a stray ESP32 AsyncTCP copy on the library search path breaks this build
+captive_portal:
+
+mqtt:
+  broker: 192.168.178.84

docker/test_configs/esp32-arduino-esp-idf.yaml

@@ -0,0 +1,10 @@
+esphome:
+  name: docker-test-esp32-ard-idf
+
+esp32:
+  variant: esp32
+  framework:
+    type: arduino
+  toolchain: esp-idf
+
+logger:

docker/test_configs/esp32-arduino-platformio.yaml

@@ -0,0 +1,10 @@
+esphome:
+  name: docker-test-esp32-ard-pio
+
+esp32:
+  variant: esp32
+  framework:
+    type: arduino
+  toolchain: platformio
+
+logger:

docker/test_configs/esp32-idf-esp-idf.yaml

@@ -0,0 +1,10 @@
+esphome:
+  name: docker-test-esp32-idf-idf
+
+esp32:
+  variant: esp32
+  framework:
+    type: esp-idf
+  toolchain: esp-idf
+
+logger:

docker/test_configs/esp32-idf-platformio.yaml

@@ -0,0 +1,10 @@
+esphome:
+  name: docker-test-esp32-idf-pio
+
+esp32:
+  variant: esp32
+  framework:
+    type: esp-idf
+  toolchain: platformio
+
+logger:

docker/test_configs/esp8266-arduino-native.yaml

@@ -0,0 +1,8 @@
+esphome:
+  name: docker-test-esp8266-native
+
+esp8266:
+  board: d1_mini
+  toolchain: arduino
+
+logger:

docker/test_configs/esp8266-arduino.yaml

@@ -0,0 +1,9 @@
+esphome:
+  name: docker-test-esp8266-arduino
+
+esp8266:
+  board: d1_mini
+  # The PlatformIO path stays covered whatever the default is
+  toolchain: platformio
+
+logger:

docker/test_configs/host.yaml

@@ -0,0 +1,6 @@
+esphome:
+  name: docker-test-host
+
+host:
+
+logger:

docker/test_configs/ln882x-arduino.yaml

@@ -0,0 +1,7 @@
+esphome:
+  name: docker-test-ln882x-arduino
+
+ln882x:
+  board: generic-ln882h
+
+logger:

docker/test_configs/nrf52.yaml

@@ -0,0 +1,8 @@
+esphome:
+  name: docker-test-nrf52
+
+nrf52:
+  board: adafruit_itsybitsy_nrf52840
+  bootloader: adafruit_nrf52_sd140_v6
+
+logger:

docker/test_configs/rp2040-arduino.yaml

@@ -0,0 +1,7 @@
+esphome:
+  name: docker-test-rp2040-arduino
+
+rp2040:
+  variant: rp2040
+
+logger:

docker/test_configs/rtl87xx-arduino.yaml

@@ -0,0 +1,7 @@
+esphome:
+  name: docker-test-rtl87xx-arduino
+
+rtl87xx:
+  board: generic-rtl8710bn-2mb-788k
+
+logger:

esphome/address_cache.py

@@ -148,6 +148,13 @@ class AddressCache:
                 continue
             hostname, ips = arg.split("=", 1)
             # Normalize hostname for consistent lookups
-            normalized = normalize_hostname(hostname)
-            cache[normalized] = [ip.strip() for ip in ips.split(",")]
+            normalized = normalize_hostname(hostname.strip())
+            addresses = [ip for value in ips.split(",") if (ip := value.strip())]
+            if not normalized or not addresses:
+                _LOGGER.warning(
+                    "Invalid cache entry: %s (hostname and at least one address are required)",
+                    arg,
+                )
+                continue
+            cache[normalized] = addresses
         return cache

esphome/analyze_memory/cli.py

@@ -20,6 +20,7 @@ from . import (
     RAM_SECTIONS,
     MemoryAnalyzer,
 )
+from .toolchain import find_elf_path, find_idedata_path, idedata_candidates
 
 if TYPE_CHECKING:
     from . import ComponentMemory
@@ -759,45 +760,25 @@ def main():
         print(f"Error: {build_path} is not a directory", file=sys.stderr)
         sys.exit(1)
 
-    # Find firmware.elf
-    elf_file = None
-    for elf_candidate in [
-        build_path / "firmware.elf",
-        build_path / ".pioenvs" / build_path.name / "firmware.elf",
-    ]:
-        if elf_candidate.exists():
-            elf_file = str(elf_candidate)
-            break
-
-    if not elf_file:
-        print(f"Error: firmware.elf not found in {build_dir}", file=sys.stderr)
+    elf_path = find_elf_path(build_path)
+    if not elf_path:
+        print(f"Error: no firmware ELF found in {build_dir}", file=sys.stderr)
         sys.exit(1)
-
-    # Find idedata.json - check current directory first, then home
-    device_name = build_path.name
-    idedata_candidates = [
-        Path.cwd() / ".esphome" / "idedata" / f"{device_name}.json",
-        Path.home() / ".esphome" / "idedata" / f"{device_name}.json",
-    ]
+    elf_file = str(elf_path)
 
     idedata = None
-    for idedata_path in idedata_candidates:
-        if not idedata_path.exists():
-            continue
+    if idedata_path := find_idedata_path(build_path):
         try:
             with idedata_path.open(encoding="utf-8") as f:
                 raw_data = json.load(f)
             idedata = IDEData(raw_data)
             print(f"Loaded idedata from: {idedata_path}", file=sys.stderr)
-            break
         except (json.JSONDecodeError, OSError) as e:
             print(f"Warning: Failed to load idedata: {e}", file=sys.stderr)
 
     if not idedata:
-        print(
-            f"Warning: idedata not found (searched {idedata_candidates[0]} and {idedata_candidates[1]})",
-            file=sys.stderr,
-        )
+        searched = "\n  ".join(str(p) for p in idedata_candidates(build_path))
+        print(f"Warning: idedata not found, searched:\n  {searched}", file=sys.stderr)
 
     analyzer = MemoryAnalyzerCLI(elf_file, idedata=idedata)
     analyzer.analyze()

esphome/analyze_memory/ram_strings.py

@@ -8,7 +8,7 @@ memory-constrained platforms like ESP8266.
 from __future__ import annotations
 
 from collections import defaultdict
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 import logging
 from pathlib import Path
 import re
@@ -65,6 +65,7 @@ class RamSymbol:
     size: int
     section: str
     demangled: str = ""  # Demangled name, set after batch demangling
+    aliases: list[str] = field(default_factory=list)  # Other names at same address
 
 
 class RamStringsAnalyzer:
@@ -235,6 +236,11 @@ class RamStringsAnalyzer:
         except (subprocess.CalledProcessError, FileNotFoundError):
             return
 
+        # Track symbols by address so aliases (multiple names for the same
+        # object, e.g. the newlib __lock___* mutexes that all alias one
+        # StaticSemaphore_t) are reported once instead of once per name.
+        symbols_by_addr: dict[int, RamSymbol] = {}
+
         for line in output.split("\n"):
             parts = line.split()
             if len(parts) < 4:
@@ -253,6 +259,18 @@ class RamStringsAnalyzer:
             if sym_type not in DATA_SYMBOL_TYPES:
                 continue
 
+            if (existing := symbols_by_addr.get(addr)) is not None:
+                # Prefer a global (uppercase type) name as the primary so
+                # nm output order can't hide it behind a local alias.
+                if sym_type.isupper() and existing.sym_type.islower():
+                    existing.aliases.append(existing.name)
+                    existing.name = name
+                    existing.sym_type = sym_type
+                else:
+                    existing.aliases.append(name)
+                existing.size = max(existing.size, size)
+                continue
+
             # Check if symbol is in a RAM section
             for section_name in self.ram_sections:
                 if section_name not in self.sections:
@@ -260,15 +278,15 @@ class RamStringsAnalyzer:
 
                 section = self.sections[section_name]
                 if section.address <= addr < section.address + section.size:
-                    self.ram_symbols.append(
-                        RamSymbol(
-                            name=name,
-                            sym_type=sym_type,
-                            address=addr,
-                            size=size,
-                            section=section_name,
-                        )
+                    symbol = RamSymbol(
+                        name=name,
+                        sym_type=sym_type,
+                        address=addr,
+                        size=size,
+                        section=section_name,
                     )
+                    symbols_by_addr[addr] = symbol
+                    self.ram_symbols.append(symbol)
                     break
 
     def _demangle_symbols(self) -> None:
@@ -436,7 +454,13 @@ class RamStringsAnalyzer:
         for symbol in largest_symbols:
             # Use demangled name if available, otherwise raw name
             display_name = symbol.demangled or symbol.name
-            name_display = display_name[:49] if len(display_name) > 49 else display_name
+            # Truncate the name, not the alias note, so merged aliases stay
+            # visible even for long demangled C++ names.
+            alias_note = f" (+{len(symbol.aliases)} aliases)" if symbol.aliases else ""
+            max_name_len = 49 - len(alias_note)
+            if len(display_name) > max_name_len:
+                display_name = display_name[:max_name_len]
+            name_display = display_name + alias_note
             lines.append(
                 f"{name_display:<50} {symbol.sym_type:<6} {symbol.size:>8} B  {symbol.section}"
             )

esphome/analyze_memory/toolchain.py

@@ -23,6 +23,82 @@ TOOLCHAIN_PREFIXES = [
 ]
 
 
+def find_elf_path(build_path: Path) -> Path | None:
+    """Locate the firmware ELF inside an ESPHome build directory.
+
+    The layout depends on the toolchain that produced the build, so try each
+    known one in turn.
+
+    Args:
+        build_path: Path to an ESPHome build directory
+
+    Returns:
+        Path to the ELF file, or None if no known layout matches
+    """
+    name = build_path.name
+    for candidate in (
+        # Native ESP-IDF: the build writes build/<name>.elf, which ESPHome copies
+        # to build/firmware.elf (see espidf.toolchain.create_elf_copy)
+        build_path / "build" / "firmware.elf",
+        # PlatformIO
+        build_path / "firmware.elf",
+        build_path / ".pioenvs" / name / "firmware.elf",
+        # LibreTiny uses raw_firmware.elf
+        build_path / "raw_firmware.elf",
+        build_path / ".pioenvs" / name / "raw_firmware.elf",
+        # Zephyr (nRF52); the SDK nests the artifacts one level deeper from 2.9.2
+        build_path / ".pioenvs" / name / "zephyr" / "zephyr" / "zephyr.elf",
+        build_path / ".pioenvs" / name / "zephyr" / "zephyr.elf",
+    ):
+        if candidate.is_file():
+            return candidate
+    return None
+
+
+def idedata_candidates(build_path: Path) -> list[Path]:
+    """Return the idedata locations searched for a build directory, in order.
+
+    Exposed so a caller reporting "not found" can name the paths it tried
+    without keeping its own copy of the list.
+
+    Args:
+        build_path: Path to an ESPHome build directory
+
+    Returns:
+        The candidate idedata JSON paths, most specific first
+    """
+    name = build_path.name
+    data_dir = build_path.parent.parent / "idedata"
+    # Native backends suffix the cache by toolchain (<name>.arduino.json)
+    suffixed = sorted(data_dir.glob(f"{name}.*.json")) if data_dir.is_dir() else []
+    return [
+        # In .pioenvs for test builds
+        build_path / ".pioenvs" / name / "idedata.json",
+        # Both toolchains cache it in the data dir, which holds this build dir:
+        # <data_dir>/idedata/<name>.json next to <data_dir>/build/<name>
+        build_path.parent.parent / "idedata" / f"{name}.json",
+        *suffixed,
+        # Regular builds, invoked from the config dir or from anywhere
+        Path.cwd() / ".esphome" / "idedata" / f"{name}.json",
+        Path.home() / ".esphome" / "idedata" / f"{name}.json",
+    ]
+
+
+def find_idedata_path(build_path: Path) -> Path | None:
+    """Locate the idedata JSON belonging to an ESPHome build directory.
+
+    Args:
+        build_path: Path to an ESPHome build directory
+
+    Returns:
+        Path to the idedata JSON, or None if it was not found
+    """
+    for candidate in idedata_candidates(build_path):
+        if candidate.is_file():
+            return candidate
+    return None
+
+
 def _find_in_platformio_packages(tool_name: str) -> str | None:
     """Search for a tool in PlatformIO package directories.
 

esphome/api_client.py

@@ -0,0 +1,196 @@
+from __future__ import annotations
+
+import asyncio
+from contextlib import suppress
+import logging
+import threading
+from typing import TYPE_CHECKING, Any
+import warnings
+
+# Suppress protobuf version warnings
+with warnings.catch_warnings():
+    warnings.filterwarnings(
+        "ignore", category=UserWarning, message=".*Protobuf gencode version.*"
+    )
+    from aioesphomeapi import APIClient, parse_log_message
+    from aioesphomeapi.log_runner import async_run
+
+from esphome.const import CONF_ENCRYPTION, CONF_KEY, CONF_PORT, __version__
+from esphome.core import CORE
+from esphome.stacktrace import LogLineProcessor
+from esphome.util import safe_print
+
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
+    from aioesphomeapi.api_pb2 import SubscribeLogsResponse  # pylint: disable=no-name-in-module
+
+
+_LOGGER = logging.getLogger(__name__)
+
+
+async def async_run_logs(
+    config: dict[str, Any],
+    addresses: list[str],
+    subscribe_states: bool = True,
+    mqtt_resolver: Callable[[threading.Event], list[str]] | None = None,
+) -> None:
+    """Run the logs command in the event loop.
+
+    If ``mqtt_resolver`` is given, it is called in a worker thread (paho-mqtt
+    has no asyncio support on Windows) concurrently with the connection
+    attempts to ``addresses``, and any addresses it discovers are fed into
+    the running client. It owns its own failure handling (returning [] when
+    discovery fails) and must honor the ``threading.Event`` it is passed so
+    teardown is not delayed by the lookup's wait window; the initial broker
+    connect itself is only bounded by the socket timeout.
+    """
+    from datetime import datetime
+
+    conf = config["api"]
+    name = config["esphome"]["name"]
+    port: int = int(conf[CONF_PORT])
+    noise_psk: str | None = None
+    if (encryption := conf.get(CONF_ENCRYPTION)) and (key := encryption.get(CONF_KEY)):
+        noise_psk = key
+
+    _LOGGER.info(
+        "Starting log output from %s using esphome API", " or ".join(addresses)
+    )
+
+    cli = APIClient(
+        addresses[0],  # Primary address for compatibility
+        port,
+        "",  # Password auth removed in 2026.1.0
+        client_info=f"ESPHome Logs {__version__}",
+        noise_psk=noise_psk,
+        addresses=addresses,  # Pass all addresses for automatic retry
+        provide_time=False,
+    )
+
+    # Decoder resolution policy lives in LogLineProcessor.
+    processor = LogLineProcessor(config, CORE.target_platform)
+
+    mqtt_task: asyncio.Task[None] | None = None
+    mqtt_stop_event = threading.Event()
+
+    def _cancel_mqtt_discovery() -> None:
+        """Stop the broker lookup once a connection has been established.
+
+        Its answer is only useful while still disconnected: after that it
+        either duplicates the connected address or arrives too late to
+        matter, so don't keep an idle broker session open for it.
+        """
+        mqtt_stop_event.set()
+        if mqtt_task is not None and not mqtt_task.done():
+            mqtt_task.cancel()
+
+    async def _resolve_mqtt_addresses() -> None:
+        """Discover the device address via the MQTT broker in the background."""
+        try:
+            mqtt_ips = await asyncio.to_thread(mqtt_resolver, mqtt_stop_event)
+            if not mqtt_ips:
+                _LOGGER.debug(
+                    "MQTT discovery %s",
+                    "aborted" if mqtt_stop_event.is_set() else "found no addresses",
+                )
+                return
+            if cli.add_addresses(mqtt_ips):
+                _LOGGER.info("Discovered address(es) via MQTT: %s", ", ".join(mqtt_ips))
+            else:
+                _LOGGER.debug(
+                    "MQTT-discovered address(es) already known: %s", ", ".join(mqtt_ips)
+                )
+        except Exception:  # pylint: disable=broad-except
+            # A background task failure would otherwise stay invisible for
+            # the whole session and only re-raise at teardown
+            _LOGGER.exception("MQTT address discovery failed")
+
+    def on_log(msg: SubscribeLogsResponse) -> None:
+        """Handle a new log message."""
+        time_ = datetime.now().astimezone()
+        message: bytes = msg.message
+        text = message.decode("utf8", "backslashreplace")
+        nanoseconds = time_.microsecond // 1000
+        timestamp = (
+            f"[{time_.hour:02}:{time_.minute:02}:{time_.second:02}.{nanoseconds:03}]"
+        )
+        for parsed_msg in parse_log_message(text, timestamp):
+            # safe_print handles the dashboard \033 escaping and falls back
+            # to backslashreplace encoding on stdouts that can't represent
+            # the wifi signal-bar block characters (Windows redirected
+            # cp1252 pipe).
+            safe_print(parsed_msg)
+        for raw_line in text.splitlines():
+            processor.process_line(raw_line)
+
+    # Safe to fall back to plaintext here only for this diagnostics use
+    # case: the stream is one-way from device to client, and this code
+    # never accepts commands or acts on any message the device sends.
+    # An on-path attacker could still both inject fabricated log lines
+    # and passively read the device's log output (and any state data
+    # delivered when subscribe_states is enabled), so this does lose
+    # confidentiality as well as authentication/integrity. That tradeoff
+    # is acceptable for operator-visible logs, which aioesphomeapi also
+    # warns may come from an unverified device. Never mirror this opt-in
+    # for any connection that sends data to the device or uses Home
+    # Assistant actions.
+    stop = await async_run(
+        cli,
+        on_log,
+        name=name,
+        subscribe_states=subscribe_states,
+        allow_plaintext_fallback=True,
+        # A top-level ``deep_sleep:`` block means the device is only awake
+        # briefly; cap the reconnect backoff so a wake window is not missed.
+        deep_sleep="deep_sleep" in config,
+        on_connect=_cancel_mqtt_discovery if mqtt_resolver is not None else None,
+    )
+    try:
+        # Don't start (or keep) the broker lookup if a connection already
+        # succeeded; the stop event doubles as the not-needed-anymore latch
+        # and get_esphome_device_ip returns immediately when it is set.
+        if mqtt_resolver is not None and not mqtt_stop_event.is_set():
+            mqtt_task = asyncio.create_task(_resolve_mqtt_addresses())
+        await asyncio.Event().wait()
+    finally:
+        try:
+            if mqtt_task is not None:
+                # Unblock the worker thread first so it can't hold up
+                # loop.shutdown_default_executor() for the full lookup timeout.
+                mqtt_stop_event.set()
+                # Give the worker a moment to exit through its own error
+                # handling; cancelling first would race out a late failure.
+                done, _ = await asyncio.wait([mqtt_task], timeout=1.0)
+                if not done:
+                    mqtt_task.cancel()
+                # return_exceptions keeps a CancelledError from the cancel()
+                # above from re-raising here and jumping over the stop() below.
+                # The task handles Exception itself, so only a BaseException
+                # escape (e.g. SystemExit from the worker) can land here.
+                (result,) = await asyncio.gather(mqtt_task, return_exceptions=True)
+                if isinstance(result, BaseException) and not isinstance(
+                    result, asyncio.CancelledError
+                ):
+                    _LOGGER.error("MQTT address discovery failed", exc_info=result)
+        finally:
+            # Must run even if a second cancellation lands mid-cleanup above
+            await stop()
+
+
+def run_logs(
+    config: dict[str, Any],
+    addresses: list[str],
+    subscribe_states: bool = True,
+    mqtt_resolver: Callable[[threading.Event], list[str]] | None = None,
+) -> None:
+    """Run the logs command."""
+    with suppress(KeyboardInterrupt):
+        asyncio.run(
+            async_run_logs(
+                config,
+                addresses,
+                subscribe_states=subscribe_states,
+                mqtt_resolver=mqtt_resolver,
+            )
+        )

esphome/arduino/library.py

@@ -0,0 +1,549 @@
+"""Arduino-core backend for the shared PlatformIO library converter.
+
+Bundled names build straight from the framework tree; everything else goes
+through ``esphome.platformio.library``. Mirrors ``lib_ldf_mode=off``: each
+library builds its own archive; all include dirs join one global path. The
+host build reuses it without a framework tree: nothing is bundled there and
+every name resolves from the registry.
+
+Deviations from PlatformIO: flat-layout libraries get the recursive default
+source filter; ``dot_a_linkage`` is honored; bundled libraries never run a
+manifest ``extraScript``; manifest ``-I`` flags join the global include path;
+``precompiled``/``ldflags`` properties are refused by name.
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+import logging
+from pathlib import Path
+import re
+
+from esphome.core import CORE, EsphomeError, Library
+from esphome.helpers import walk_files
+from esphome.platformio.extra_script import apply_extra_script
+from esphome.platformio.library import (
+    DEFAULT_BUILD_INCLUDE_DIR,
+    DEFAULT_BUILD_SRC_FILTER,
+    ESPHOME_DATA_KEY,
+    ESPHOME_DATA_LINK_FLAGS_KEY,
+    LIBRARY_HEADER_SUFFIXES,
+    SRC_FILE_EXTENSIONS,
+    ConvertedLibrary,
+    IncompatiblePlatform,
+    InvalidLibrary,
+    LibraryBackend,
+    _url_or_none,
+    check_library_data,
+    collect_filtered_files,
+    convert_libraries,
+    ensure_list,
+    is_lib_ignored,
+    lex_build_flags,
+    lib_ignore_set,
+    normalize_dependencies,
+    parse_library_json,
+    parse_library_properties,
+    warn_properties_depends,
+)
+
+_LOGGER = logging.getLogger(__name__)
+
+
+@dataclass
+class ArduinoLibrary:
+    """One resolved library, ready for the ninja generator."""
+
+    name: str
+    sources: list[Path] = field(default_factory=list)
+    include_dirs: list[Path] = field(default_factory=list)
+    # Extra compile flags private to this library's own sources
+    flags: list[str] = field(default_factory=list)
+    # PlatformIO's build.libArchive / Arduino's dot_a_linkage: when False the
+    # objects go to the linker directly (symbols nothing references survive)
+    lib_archive: bool = True
+    # Link inputs the library contributes (-L dirs / -l libs, e.g. from
+    # precompiled vendor blobs) and -Wl, options for the firmware link
+    link_dirs: list[Path] = field(default_factory=list)
+    link_libs: list[str] = field(default_factory=list)
+    link_flags: list[str] = field(default_factory=list)
+
+
+# Source-like suffixes the case-sensitive suffix map rejects
+_UNMAPPED_SOURCE_SUFFIXES = frozenset(
+    {s.lower() for s in SRC_FILE_EXTENSIONS} | {".ino"}
+)
+
+# Filename-plain names: an allowlist excludes separators, drive colons,
+# and dot-only names by shape
+_SAFE_LIBRARY_NAME_RE = re.compile(r"[A-Za-z0-9_][A-Za-z0-9_. +-]*\Z")
+
+
+def _is_safe_library_name(name: object) -> bool:
+    """Whether a name may be joined under the framework's libraries dir."""
+    return isinstance(name, str) and _SAFE_LIBRARY_NAME_RE.fullmatch(name) is not None
+
+
+def _manifest_build(name: str, data: object) -> dict:
+    """The manifest's ``build`` section; malformed manifests fail by name."""
+    build = data.get("build", {}) if isinstance(data, dict) else None
+    if not isinstance(build, dict):
+        raise EsphomeError(f"Library {name} has a malformed manifest")
+    return build
+
+
+def _resolve_src_dir(name: str, read_path: Path, build: dict) -> str:
+    """Resolve PIO's source dir: manifest srcDir, else src/Src, else the root."""
+    if "srcDir" not in build:
+        return next((d for d in ("src", "Src") if (read_path / d).is_dir()), ".")
+    # A declared srcDir (falsy included) that does not resolve is a manifest error
+    src_dir = build["srcDir"]
+    if not (isinstance(src_dir, str) and src_dir and (read_path / src_dir).is_dir()):
+        raise EsphomeError(
+            f"Library {name} declares srcDir {src_dir!r} which does not exist"
+        )
+    return src_dir
+
+
+def _reject_unsupported_link_fields(name: str, data: dict) -> None:
+    # PIO honors these; ignoring them would fail at link with no stated
+    # cause. Property values are strings, so "false" is not a declaration.
+    precompiled = data.get("precompiled")
+    if precompiled and str(precompiled).strip().lower() != "false":
+        raise EsphomeError(
+            f"Library {name} declares precompiled, which this backend does not support"
+        )
+    if data.get("ldflags"):
+        raise EsphomeError(
+            f"Library {name} declares ldflags, which this backend does not support"
+        )
+
+
+def _resolve_lib_archive(name: str, data: dict, build: dict) -> bool:
+    """build.libArchive, else dot_a_linkage (an Arduino IDE property PIO
+    ignores; a deliberate extra), else archive."""
+
+    # Strict parse: bool("false") is True
+    def _parse(key: str, raw: object) -> bool:
+        if isinstance(raw, bool):
+            return raw
+        value = str(raw).strip().lower()
+        if value in ("true", "false"):
+            return value == "true"
+        raise EsphomeError(f"Library {name} has a malformed {key} value {raw!r}")
+
+    if "libArchive" in build:
+        return _parse("libArchive", build["libArchive"])
+    if "dot_a_linkage" in data:
+        return _parse("dot_a_linkage", data["dot_a_linkage"])
+    return True
+
+
+def _classify_build_flags(
+    name: str, read_path: Path, lib: ArduinoLibrary, flag_tokens: list[str]
+) -> list[str]:
+    """Route the lexed build.flags into the library's flag lists.
+
+    Returns the ``-I`` arguments for the include-dir resolution.
+    """
+    include_flags: list[str] = []
+    for tok in flag_tokens:
+        if tok.startswith("-I"):
+            include_flags.append(tok[2:])
+        elif tok.startswith("-L"):
+            link_dir = (read_path / tok[2:]).resolve()
+            if not link_dir.is_dir():
+                # Kept (the linker ignores missing -L dirs); the warning
+                # names the culprit before a bare "cannot find -lfoo"
+                _LOGGER.warning(
+                    "Library %s declares library dir %s which does not exist",
+                    name,
+                    tok[2:],
+                )
+            lib.link_dirs.append(link_dir)
+        elif tok.startswith("-l"):
+            lib.link_libs.append(tok[2:])
+        elif tok.startswith("-Wl,"):
+            lib.link_flags.append(tok)
+        else:
+            lib.flags.append(tok)
+    return include_flags
+
+
+def _resolve_include_dirs(
+    name: str,
+    read_path: Path,
+    lib: ArduinoLibrary,
+    build: dict,
+    src_dir: str,
+    include_flags: list[str],
+) -> None:
+    include_dir = build.get("includeDir", DEFAULT_BUILD_INCLUDE_DIR)
+    if not isinstance(include_dir, str):
+        raise EsphomeError(f"Library {name} has a malformed includeDir")
+    for d, explicit in [
+        (include_dir, "includeDir" in build),
+        (src_dir, False),  # _resolve_src_dir already validated it
+        *((flag, True) for flag in include_flags),
+    ]:
+        if (path := (read_path / d)).is_dir():
+            lib.include_dirs.append(path.resolve())
+        elif explicit:
+            # Warn-and-drop (unlike srcDir): a missing include dir is
+            # harmless until a header is needed, and the compile names it
+            _LOGGER.warning(
+                "Library %s declares include dir %s which does not exist", name, d
+            )
+
+
+def _collect_lib_sources(
+    name: str,
+    read_path: Path,
+    lib: ArduinoLibrary,
+    src_dir: str,
+    src_filter: list[str],
+) -> None:
+    sources: list[Path] = []
+    dropped: list[str] = []
+    saw_header = False
+    for f in collect_filtered_files(read_path / src_dir, src_filter):
+        path = Path(f)
+        suffix = path.suffix
+        if suffix in SRC_FILE_EXTENSIONS:
+            # resolve() per file: srcFilter patterns may escape src_dir
+            sources.append(path.resolve())
+        elif suffix.lower() in _UNMAPPED_SOURCE_SUFFIXES:
+            # A source-like suffix the case-sensitive map rejects (.CPP,
+            # .ino) is a dropped compilation unit; headers fall through
+            dropped.append(path.name)
+        elif suffix.lower() in LIBRARY_HEADER_SUFFIXES:
+            saw_header = True
+    lib.sources = sorted(sources)
+    if dropped:
+        _LOGGER.warning(
+            "Library %s: %d file(s) with unmapped source suffixes are not compiled: %s",
+            name,
+            len(dropped),
+            ", ".join(sorted(dropped)),
+        )
+    if not lib.sources and not saw_header:
+        # Matched headers mean header-only; a filter matching nothing is
+        # a manifest/tree problem (a truly empty tree raises elsewhere)
+        _LOGGER.warning("Library %s: no source files matched", name)
+
+
+def _library_info(name: str, read_path: Path, data: dict) -> ArduinoLibrary:
+    """Resolve one library's sources, include dirs, and flags (PIO semantics)."""
+    build = _manifest_build(name, data)
+    _reject_unsupported_link_fields(name, data)
+    src_dir = _resolve_src_dir(name, read_path, build)
+    src_filter = ensure_list(build.get("srcFilter", DEFAULT_BUILD_SRC_FILTER))
+    if not all(isinstance(entry, str) for entry in src_filter):
+        raise EsphomeError(f"Library {name} has a malformed srcFilter")
+    lib = ArduinoLibrary(name=name, lib_archive=_resolve_lib_archive(name, data, build))
+    # PlatformIO shell-lexes each build.flags entry
+    include_flags = _classify_build_flags(
+        name, read_path, lib, lex_build_flags(build.get("flags", []), f"library {name}")
+    )
+    _resolve_include_dirs(name, read_path, lib, build, src_dir, include_flags)
+    _collect_lib_sources(name, read_path, lib, src_dir, src_filter)
+    return lib
+
+
+def _bundled_library(framework_path: Path, name: str) -> ArduinoLibrary:
+    """A library bundled with the Arduino core, read from the framework tree.
+
+    ``library.json`` wins over ``library.properties`` when both exist, as in
+    PlatformIO's LibBuilderFactory; only the JSON manifest can carry a
+    ``build`` section (srcDir, srcFilter, flags).
+    """
+    lib_dir = framework_path / "libraries" / name
+    manifest_json = lib_dir / "library.json"
+    if manifest_json.is_file():
+        try:
+            data = parse_library_json(manifest_json)
+        except ValueError as err:  # JSONDecodeError
+            raise EsphomeError(
+                f"Bundled library {name} has a corrupt library.json ({err}); "
+                "the framework install may be incomplete (run 'esphome clean-all')"
+            ) from err
+    elif (manifest := lib_dir / "library.properties").is_file():
+        data = parse_library_properties(manifest)
+    else:
+        # Debug, not warning: the legacy manifest-less layout is legal and
+        # the 3.1.2 core ships one such library (FSTools), so a warning
+        # would be unactionable noise on every build using it
+        _LOGGER.debug("Bundled library %s has no manifest; using defaults", name)
+        data = {}
+    if isinstance(data, dict):
+        # Bundled manifest deps are never walked; make the skip visible
+        if data.get("dependencies"):
+            _LOGGER.warning(
+                "Bundled library %s declares dependencies, which are not "
+                "resolved automatically; add them with add_library() if needed",
+                name,
+            )
+        warn_properties_depends(name, data)
+        build = data.get("build")
+        if isinstance(build, dict) and build.get("extraScript"):
+            # Scripts only run on the converted path; building without
+            # the script's flags would miscompile
+            raise EsphomeError(
+                f"Bundled library {name} declares an extraScript, which is "
+                "not run for bundled libraries"
+            )
+    lib = _library_info(name, lib_dir, data)
+    _assert_tree_has_code(
+        name,
+        lib_dir,
+        "the framework install may be incomplete (run 'esphome clean-all')",
+    )
+    return lib
+
+
+def _assert_tree_has_code(name: str, root: Path, hint: str) -> None:
+    """An empty or half-extracted tree can never link; fail by name (a
+    warning would scroll away and resurface as undefined symbols)."""
+    if not any(
+        Path(p).suffix in SRC_FILE_EXTENSIONS
+        or Path(p).suffix.lower() in LIBRARY_HEADER_SUFFIXES
+        for p in walk_files(root)
+    ):
+        raise EsphomeError(f"Library {name} has no sources or headers; {hint}")
+
+
+def _external_short_name(name: str) -> str:
+    """The short library name of a requested spec.
+
+    "owner/Name" and plain names take the last path segment; "Name=<url>"
+    takes the declared name. Git tails (".git", "#ref") are stripped like
+    the walk's URL normalization; the comparand is a manifest dependency
+    name, never a spec.
+    """
+    head, sep, tail = name.partition("=")
+    if sep and "://" in tail:
+        return head
+    short = name.rsplit("/", maxsplit=1)[-1]
+    return short.partition("#")[0].removesuffix(".git")
+
+
+def _check_unfulfilled_provides(
+    provided_requests: set[str], satisfied: set[str], still_requested: set[str]
+) -> None:
+    """Fail by name when a walk-skipped dependency was never added.
+
+    An unfulfilled provides() promise only surfaces as undefined symbols
+    at link. The walk records across re-resolutions, so a name no final
+    manifest still requests is stale state, never a failure.
+    """
+    if missing := sorted((provided_requests & still_requested) - satisfied):
+        raise EsphomeError(
+            "provides() skipped these dependencies but nothing added them: "
+            f"{', '.join(missing)}; the build is missing libraries"
+        )
+
+
+def resolve_libraries(
+    framework_path: Path | None,
+    *,
+    pio_platform: str,
+    board_mcu: str,
+    cache_key: str,
+    framework: str | None = "arduino",
+    manifest_optional: bool = False,
+) -> list[ArduinoLibrary]:
+    """Resolve every ``cg.add_library()`` entry into an :class:`ArduinoLibrary`.
+
+    ``pio_platform``/``board_mcu`` filter manifests the way PlatformIO would
+    for that core (e.g. ``espressif8266``/``esp8266``); ``cache_key`` keys the
+    shared converter's download cache. ``framework`` is the manifest
+    framework token the compatibility check warns about; None skips it.
+    A None ``framework_path`` means no core-bundled libraries exist (the
+    host build): every name resolves from the registry.
+    ``manifest_optional`` accepts libraries without a manifest, built with
+    PlatformIO's default layout.
+
+    The returned list is not topologically sorted, so the caller must link
+    the archives inside one ``--start-group``/``--end-group`` pair (the
+    bundled-first grouping is incidental).
+    """
+    bundled: list[ArduinoLibrary] = []
+    external: list[Library] = []
+    # PlatformIO's lib_ignore covers framework-bundled libraries too; the
+    # shared converter only filters the registry/git ones.
+    lib_ignore = lib_ignore_set()
+    bundled_dir_names: frozenset[str] = frozenset()
+    if framework_path is not None:
+        # Exact directory names keep membership case-sensitive everywhere
+        # (an is_dir() probe would match "wire" on macOS/Windows and build
+        # the bundled Wire twice)
+        libraries_dir = framework_path / "libraries"
+        if not libraries_dir.is_dir():
+            # A registry fallback would fail later with a misleading
+            # package-not-found error per bundled name
+            raise EsphomeError(
+                f"{libraries_dir} is missing; the framework install may be "
+                "incomplete (run 'esphome clean-all')"
+            )
+        bundled_dir_names = frozenset(
+            p.name for p in libraries_dir.iterdir() if p.is_dir()
+        )
+
+    def _provided(name: object) -> bool:
+        return _is_safe_library_name(name) and name in bundled_dir_names
+
+    for library in CORE.platformio_libraries.values():
+        if is_lib_ignored(library.name, lib_ignore):
+            continue
+        # Bundled only for a bare name with a matching framework dir; pinned
+        # or unmatched names resolve from the registry, as under PlatformIO.
+        if not library.repository and not library.version and _provided(library.name):
+            # Bundled manifest deps are not walked; _bundled_library warns
+            bundled.append(_bundled_library(framework_path, library.name))
+        else:
+            external.append(library)
+
+    converted: list[ArduinoLibrary] = []
+    bundled_names = {lib.name for lib in bundled}
+    converted_manifest_names: set[str] = set()
+    # Bundled candidates skipped on purpose (platform filter); the
+    # provides() reconciliation must count them as satisfied
+    knowingly_skipped: set[str] = set()
+    # Dependency names of the manifests actually emitted; a walk recording
+    # for a since-re-resolved manifest must not fail the reconciliation
+    final_dep_names: set[str] = set()
+    # Ordered set of bundled dependency names to add once conversion is done
+    pending_bundled: dict[str, None] = {}
+    # Deps matching a separately-requested external are already in the build
+    # (a duplicate archive means duplicate-symbol link errors)
+    external_short_names = {
+        _external_short_name(lib.name) for lib in external if lib.name
+    }
+
+    def _add_bundled_dependencies(component: ConvertedLibrary) -> None:
+        # A version-less bare name ("Hash") is a core-bundled library the
+        # shared converter cannot resolve from the registry
+        for dep in normalize_dependencies(
+            component.data.get("dependencies"), component.name
+        ):
+            # normalize_dependencies guarantees a non-empty str name
+            name = dep["name"]
+            final_dep_names.add(name)
+            if "/" in name:
+                owner, _, pkg = name.partition("/")
+                if _is_safe_library_name(owner) and _is_safe_library_name(pkg):
+                    # Owner-qualified; the converter resolves it from the registry
+                    continue
+            if not _is_safe_library_name(name):
+                # The name becomes a path component; never join a traversal
+                _LOGGER.warning(
+                    "Ignoring malformed dependency entry %r of library %s",
+                    dep,
+                    component.name,
+                )
+                continue
+            if name in external_short_names:
+                if _provided(name):
+                    # A bundled copy is suppressed; a coincidental name
+                    # collision would surface as link errors
+                    _LOGGER.warning(
+                        "Dependency %s of %s is assumed satisfied by a "
+                        "requested external library; the bundled copy is "
+                        "not added",
+                        name,
+                        component.name,
+                    )
+                else:
+                    _LOGGER.debug(
+                        "Dependency %s of %s assumed satisfied by a requested "
+                        "external library",
+                        name,
+                        component.name,
+                    )
+                continue
+            if name in bundled_names or is_lib_ignored(name, lib_ignore):
+                continue
+            if _url_or_none(dep.get("version")) is not None:
+                # A URL names one specific source; never add the bundled copy
+                continue
+            if dep.get("owner") or not _provided(name):
+                # Only owner-less framework-tree names take the bundled
+                # copy (PIO's process_dependencies); the walk reports drops
+                continue
+            try:
+                # framework=None: the walk already warned for non-platform
+                # causes; debug keeps one fault from warning twice (pinned
+                # by test_nonplatform_rejection_warns_once_through_real_converter)
+                check_library_data(dep, pio_platform, None)
+            except IncompatiblePlatform as err:
+                # A knowing skip (platform filter), not a broken promise
+                knowingly_skipped.add(name)
+                _LOGGER.debug("Skip bundled candidate %s: %s", name, err)
+                continue
+            except InvalidLibrary as err:
+                # Malformed manifest data never counts as satisfied; the
+                # walk owns the warning (see the warns-once test above)
+                _LOGGER.debug("Skip malformed bundled candidate %s: %s", name, err)
+                continue
+            # Deferred: a later manifest name may satisfy this
+            pending_bundled.setdefault(name)
+
+    def _emit(component: ConvertedLibrary) -> None:
+        apply_extra_script(
+            component, board_mcu=lambda: board_mcu, pio_platform=pio_platform
+        )
+        _assert_tree_has_code(
+            component.get_require_name(),
+            component.source_dir,
+            "the download may be incomplete (run 'esphome clean-all')",
+        )
+        if isinstance(manifest_name := component.data.get("name"), str):
+            converted_manifest_names.add(manifest_name)
+        lib = _library_info(
+            component.get_require_name(), component.source_dir, component.data
+        )
+        # Extra-script LINKFLAGS travel outside build.flags; dropping
+        # them would link wrong with no stated cause
+        lib.link_flags.extend(
+            component.data.get(ESPHOME_DATA_KEY, {}).get(
+                ESPHOME_DATA_LINK_FLAGS_KEY, []
+            )
+        )
+        converted.append(lib)
+        _add_bundled_dependencies(component)
+
+    backend = LibraryBackend(
+        platform=pio_platform,
+        framework=framework,
+        emit=_emit,
+        cache_key=cache_key,
+        # The walk must not resolve bundled names from the registry;
+        # _add_bundled_dependencies adds them after emit
+        provides=_provided,
+        manifest_optional=manifest_optional,
+    )
+    if external:
+        convert_libraries(external, backend)
+    for name in pending_bundled:
+        if name in converted_manifest_names:
+            # The converted library is this one; the bundled copy would
+            # double the archive. Warn like the external_short_names twin.
+            _LOGGER.warning(
+                "Dependency %s is assumed satisfied by a converted library's "
+                "manifest name; the bundled copy is not added",
+                name,
+            )
+            continue
+        bundled_names.add(name)
+        bundled.append(_bundled_library(framework_path, name))
+
+    _check_unfulfilled_provides(
+        backend.provided_requests,
+        bundled_names
+        | converted_manifest_names
+        | external_short_names
+        | knowingly_skipped,
+        final_dep_names,
+    )
+
+    return bundled + converted

esphome/arduino8266/__init__.py

@@ -0,0 +1,9 @@
+"""Native (PlatformIO-free) build support for the ESP8266 Arduino core.
+
+This package downloads the Arduino ESP8266 core and the xtensa-lx106
+toolchain, generates a ninja build for them plus the ESPHome sources, and
+drives the build directly — the ESP8266 equivalent of ``esphome.espidf``.
+
+Deliberately importable without the esp8266 component to avoid circular
+imports; the component wires these modules in via lazy imports.
+"""

esphome/arduino8266/framework.py

@@ -0,0 +1,207 @@
+"""Download and install the Arduino ESP8266 core, toolchain, and ninja.
+
+Artifacts land in a machine-global cache (shared across projects, like the
+ESP-IDF install in ``esphome.espidf.framework``):
+
+    <cache>/arduino8266/frameworks/<tag>/   the Arduino core
+    <cache>/arduino8266/toolchains/<tag>/   xtensa-lx106-elf gcc 10.3
+
+Both come from esphome-libs releases pinned below;
+``ESPHOME_ARDUINO8266_*_MIRRORS`` overrides the URLs, with ``{VERSION}``
+standing for the release tag. ninja comes from PATH or the ninja PyPI wheel.
+"""
+
+from __future__ import annotations
+
+import os
+from pathlib import Path
+from typing import NamedTuple
+
+from esphome.build_helpers.ccache import ccache_env
+from esphome.build_helpers.ninja import find_ninja
+from esphome.build_helpers.tools_cache import ARDUINO8266_TOOLS_CACHE, tools_cache_path
+from esphome.core import EsphomeError, Version
+from esphome.framework_helpers import str_to_lst_of_str
+from esphome.platformio.registry import (
+    Download,
+    PackageSpec,
+    Resolver,
+    get_systype,
+    install_packages,
+    prefetch_packages,
+)
+
+FRAMEWORK_PACKAGE = "arduino-esp8266"
+_FRAMEWORK_RELEASES = "https://github.com/esphome-libs/arduino-esp8266/releases/"
+
+
+class FrameworkRelease(NamedTuple):
+    tag: str
+    sha256: str
+    size: int
+
+    def download(self) -> Download:
+        archive = f"{FRAMEWORK_PACKAGE}-{self.tag}.tar.gz"
+        url = f"{_FRAMEWORK_RELEASES}download/{self.tag}/{archive}"
+        return Download(url, self.sha256, self.size)
+
+
+# Arduino core version -> its build in esphome-libs/arduino-esp8266
+FRAMEWORK_RELEASES: dict[Version, FrameworkRelease] = {
+    Version(3, 1, 2): FrameworkRelease(
+        "3.1.2-esphome.3",
+        "3af533a85163804e164500a5b254dc23b922080e25ecb7dc37b3dd8ac84651db",
+        37185650,
+    ),
+}
+
+TOOLCHAIN_PACKAGE = "toolchain-xtensa-lx106-elf"
+# gcc 10.3, the toolchain Arduino core 3.x builds with; the build
+# generator's compile flags are tuned to it.
+TOOLCHAIN_VERSION = "10.3.0-esphome.3"
+_TOOLCHAIN_RELEASES = (
+    "https://github.com/esphome-libs/xtensa-lx106-elf-toolchain/releases/"
+)
+# Registry system tag -> (sha256, size) of that host's archive
+TOOLCHAIN_BUILDS: dict[str, tuple[str, int]] = {
+    "darwin_arm64": (
+        "ad0ea929238d9b527c43e0d20257acc2a5680bdb0d2b1cf8f2c99169786217e9",
+        60747260,
+    ),
+    "darwin_x86_64": (
+        "997090cac80c8eb59604ec52b73fb3568e5f2d9b2971f83c3f2a5c545c9c949c",
+        64065752,
+    ),
+    "linux_aarch64": (
+        "403f14d0d5755682fcd6eaea1e4c457a1d8cad30690e52fb0c804d8db266b908",
+        67487067,
+    ),
+    "linux_x86_64": (
+        "64f21fa3ba736b363dc8c234dbbb39e005e574e33de08a78e9a00c55a64ae706",
+        68343668,
+    ),
+    "windows_amd64": (
+        "1e62abf8df041ce5158f5770dffef955f2f981ce7f7adee88fcd5a8197705aa6",
+        67536405,
+    ),
+}
+
+ESPHOME_ARDUINO8266_FRAMEWORK_MIRRORS = str_to_lst_of_str(
+    os.environ.get("ESPHOME_ARDUINO8266_FRAMEWORK_MIRRORS", "")
+)
+ESPHOME_ARDUINO8266_TOOLCHAIN_MIRRORS = str_to_lst_of_str(
+    os.environ.get("ESPHOME_ARDUINO8266_TOOLCHAIN_MIRRORS", "")
+)
+
+
+def get_arduino8266_tools_path() -> Path:
+    # Machine-global so all projects share one install; see
+    # espidf.framework.get_idf_tools_path for the location rationale.
+    return tools_cache_path(*ARDUINO8266_TOOLS_CACHE)
+
+
+def framework_release(version: Version) -> FrameworkRelease:
+    if (release := FRAMEWORK_RELEASES.get(version)) is None:
+        raise EsphomeError(
+            f"'toolchain: arduino' has no build of Arduino core {version}; "
+            f"available: {', '.join(str(v) for v in FRAMEWORK_RELEASES)}. "
+            "Use one of those or 'toolchain: platformio'"
+        )
+    return release
+
+
+def get_framework_path(tag: str) -> Path:
+    return get_arduino8266_tools_path() / "frameworks" / tag
+
+
+def get_toolchain_path() -> Path:
+    return get_arduino8266_tools_path() / "toolchains" / TOOLCHAIN_VERSION
+
+
+def toolchain_download() -> Download:
+    """The toolchain archive for the current host."""
+    systype = get_systype()
+    if (build := TOOLCHAIN_BUILDS.get(systype)) is None:
+        raise EsphomeError(
+            f"There is no ESP8266 toolchain for this system ({systype}); "
+            f"supported systems are {', '.join(sorted(TOOLCHAIN_BUILDS))}. "
+            "Either set 'toolchain: platformio' under 'esp8266:', or point "
+            "ESPHOME_ARDUINO8266_TOOLCHAIN_MIRRORS at a toolchain archive"
+        )
+    sha256, size = build
+    archive = f"{TOOLCHAIN_PACKAGE}-{TOOLCHAIN_VERSION}-{systype}.tar.gz"
+    url = f"{_TOOLCHAIN_RELEASES}download/{TOOLCHAIN_VERSION}/{archive}"
+    return Download(url, sha256, size)
+
+
+class InstalledPaths(NamedTuple):
+    """Locations of the installed framework, toolchain, and ninja binary."""
+
+    framework: Path
+    toolchain: Path
+    ninja: Path
+
+
+def check_and_install(framework_version: Version) -> InstalledPaths:
+    """Ensure framework, toolchain, and ninja are installed; return their paths."""
+    release = framework_release(framework_version)
+    # Probe the cheap local dependency before ~110 MB of downloads
+    ninja_path = find_ninja()
+    framework_path = get_framework_path(release.tag)
+    downloads_dir = get_arduino8266_tools_path() / "downloads"
+    toolchain_path = get_toolchain_path()
+    # One spec per package: the prefetch and the installs must agree
+    specs = (
+        PackageSpec(
+            FRAMEWORK_PACKAGE,
+            release.tag,
+            framework_path,
+            ESPHOME_ARDUINO8266_FRAMEWORK_MIRRORS,
+            ("cores/esp8266", "tools/sdk", "libraries"),
+        ),
+        PackageSpec(
+            TOOLCHAIN_PACKAGE,
+            TOOLCHAIN_VERSION,
+            toolchain_path,
+            ESPHOME_ARDUINO8266_TOOLCHAIN_MIRRORS,
+            # xtensa-lx106-elf pins the target: every gcc package has a bin/
+            ("bin", "xtensa-lx106-elf"),
+        ),
+    )
+    # Resolved only when a download is needed, so an installed toolchain
+    # keeps working on a host without a build; a mirror override wins
+    resolvers: dict[str, Resolver] = {}
+    if not ESPHOME_ARDUINO8266_FRAMEWORK_MIRRORS:
+        resolvers[FRAMEWORK_PACKAGE] = release.download
+    if not ESPHOME_ARDUINO8266_TOOLCHAIN_MIRRORS:
+        resolvers[TOOLCHAIN_PACKAGE] = toolchain_download
+    # Fetch both archives at once; the installs verify and extract them.
+    # One spec list for both, so the two phases cannot drift.
+    prefetch_packages(specs, downloads_dir, resolvers)
+    install_packages(specs, downloads_dir, resolvers)
+    return InstalledPaths(
+        framework=framework_path, toolchain=toolchain_path, ninja=ninja_path
+    )
+
+
+def toolchain_tool(toolchain_path: Path, name: str) -> Path:
+    """Path to one toolchain tool (gcc, g++, ar, size, addr2line, ...).
+
+    The single owner of the ``bin/xtensa-lx106-elf-<name>`` layout and the
+    Windows suffix, so a toolchain package bump touches one spot.
+    """
+    suffix = ".exe" if os.name == "nt" else ""
+    return toolchain_path / "bin" / f"xtensa-lx106-elf-{name}{suffix}"
+
+
+def get_build_env(toolchain_path: Path, ccache: str | None) -> dict[str, str]:
+    env = os.environ.copy()
+    # Drop empty entries: a trailing separator from an absent PATH would
+    # make the shell search the current directory for tools
+    parts = [
+        str(toolchain_path / "bin"),
+        *filter(None, env.get("PATH", "").split(os.pathsep)),
+    ]
+    env["PATH"] = os.pathsep.join(parts)
+    env.update(ccache_env(ccache, ARDUINO8266_TOOLS_CACHE))
+    return env