Merge remote-tracking branch 'upstream/master' into chore/plenticore-caveat

Author: iseeberg79

.coderabbit.yaml

@@ -0,0 +1,21 @@
+# yaml-language-server: $schema=https://storage.googleapis.com/coderabbit_public_assets/schema.v2.json
+reviews:
+  # Leave the PR description untouched (no "Summary by CodeRabbit").
+  high_level_summary: false
+
+  # Suppress the walkthrough comment by disabling all its parts.
+  changed_files_summary: false
+  sequence_diagrams: false
+  estimate_code_review_effort: false
+  assess_linked_issues: false
+  related_issues: false
+  related_prs: false
+  poem: false
+
+  # Auto-apply suggested labels and reviewers to the PR.
+  auto_apply_labels: true
+  auto_assign_reviewers: true
+
+  # Flag spam / low-quality PRs (public GitHub repos only).
+  slop_detection:
+    enabled: true

.github/workflows/claude-code-review-run.yml

@@ -0,0 +1,137 @@
+name: Claude Code Review (run)
+
+# Stage 2 (privileged): triggered when the "Claude Code Review" workflow above
+# completes. workflow_run executes in the base-repo context, so it has access
+# to secrets and id-token even for PRs from forks. This is what makes reviewing
+# external contributors' PRs possible.
+#
+# Security: the triggering workflow file is fork-controlled, so the uploaded
+# artifact is untrusted. We validate the PR number is purely numeric before
+# using it, and we only check out the base ref (never the PR head into the
+# workspace root) so untrusted code is never executed here. See
+# https://securitylab.github.com/research/github-actions-preventing-pwn-requests/
+on:
+  workflow_run:
+    workflows: ["Claude Code Review"]
+    types: [completed]
+
+jobs:
+  gate:
+    # Only react to PR runs that finished successfully.
+    if: >
+      github.event.workflow_run.event == 'pull_request' &&
+      github.event.workflow_run.conclusion == 'success'
+    runs-on: ubuntu-latest
+    permissions:
+      actions: read
+    outputs:
+      number: ${{ steps.read.outputs.number }}
+      found: ${{ steps.check.outputs.found }}
+    steps:
+      - name: Check for PR-number artifact
+        id: check
+        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v7
+        with:
+          script: |
+            const { data } = await github.rest.actions.listWorkflowRunArtifacts({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              run_id: context.payload.workflow_run.id,
+            });
+            core.setOutput('found', data.artifacts.some(a => a.name === 'pr-number') ? 'true' : 'false');
+
+      - name: Download PR number
+        if: steps.check.outputs.found == 'true'
+        uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v4
+        with:
+          name: pr-number
+          run-id: ${{ github.event.workflow_run.id }}
+          github-token: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Read and validate PR number
+        id: read
+        if: steps.check.outputs.found == 'true'
+        run: |
+          num="$(cat pr-number.txt)"
+          # Artifact comes from the fork-controlled trigger workflow — never trust it blindly.
+          if ! [[ "$num" =~ ^[0-9]+$ ]]; then
+            echo "Refusing to proceed: PR number is not numeric ($num)" >&2
+            exit 1
+          fi
+          echo "number=$num" >> "$GITHUB_OUTPUT"
+
+  claude-review:
+    needs: gate
+    if: needs.gate.outputs.found == 'true'
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      pull-requests: write # post the review on the PR
+      issues: read
+      id-token: write
+      statuses: write # mirror the review result onto the PR head commit
+    concurrency:
+      group: claude-review-run-${{ needs.gate.outputs.number }}
+      cancel-in-progress: true
+    steps:
+      # workflow_run runs are detached and never show up in the PR's checks, so
+      # publish a commit status on the PR head SHA. This surfaces a "Claude Code
+      # Review" check next to the regular CI jobs.
+      - name: Report review pending
+        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v7
+        with:
+          script: |
+            await github.rest.repos.createCommitStatus({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              sha: context.payload.workflow_run.head_sha,
+              context: 'Claude Code Review',
+              state: 'pending',
+              description: 'Review in progress',
+              target_url: `${context.serverUrl}/${context.repo.owner}/${context.repo.repo}/actions/runs/${context.runId}`,
+            });
+
+      # Default checkout (base ref) only — do NOT check out the untrusted PR head.
+      - name: Checkout repository
+        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v4
+        with:
+          fetch-depth: 1
+
+      - name: Run Claude Code Review
+        id: review
+        continue-on-error: true # always report status below, even on failure
+        uses: anthropics/claude-code-action@v1
+        with:
+          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
+          # Provided explicitly so the review can post on fork PRs; also required
+          # for allowed_non_write_users to take effect.
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          # Review PRs from external contributors who lack write access. The diff
+          # is fetched read-only and never executed; the subprocess env is scrubbed.
+          allowed_non_write_users: "*"
+          # review dependency bumps opened by Dependabot
+          allowed_bots: "dependabot[bot]"
+          plugin_marketplaces: "https://github.com/anthropics/claude-code.git"
+          plugins: "code-review@claude-code-plugins"
+          prompt: "/code-review:code-review ${{ github.repository }}/pull/${{ needs.gate.outputs.number }}"
+          # See https://github.com/anthropics/claude-code-action/blob/main/docs/usage.md
+          # or https://code.claude.com/docs/en/cli-reference for available options
+
+      - name: Report review result
+        if: always()
+        env:
+          OUTCOME: ${{ steps.review.outcome }}
+        uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v7
+        with:
+          script: |
+            const ok = process.env.OUTCOME === 'success';
+            await github.rest.repos.createCommitStatus({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              sha: context.payload.workflow_run.head_sha,
+              context: 'Claude Code Review',
+              state: ok ? 'success' : 'failure',
+              description: ok ? 'Review complete' : 'Review failed',
+              target_url: `${context.serverUrl}/${context.repo.owner}/${context.repo.repo}/actions/runs/${context.runId}`,
+            });
+            if (!ok) core.setFailed('Claude Code Review failed');

.github/workflows/claude-code-review.yml

@@ -0,0 +1,32 @@
+name: Claude Code Review
+
+# Stage 1 (unprivileged): runs on every PR, including forks. Fork PRs get a
+# read-only token and no secrets, so the review itself cannot run here. This
+# job only records the PR number; the privileged "Claude Code Review (run)"
+# workflow picks it up via workflow_run and performs the actual review.
+on:
+  pull_request:
+    types: [opened, synchronize, ready_for_review, reopened]
+
+concurrency:
+  group: claude-review-${{ github.event.pull_request.number }}
+  cancel-in-progress: true
+
+jobs:
+  prepare:
+    # Skip draft PRs
+    if: github.event.pull_request.draft == false
+    runs-on: ubuntu-latest
+    permissions: {}
+    steps:
+      - name: Save PR number
+        env:
+          PR_NUMBER: ${{ github.event.pull_request.number }}
+        run: echo "$PR_NUMBER" > pr-number.txt
+
+      - name: Upload PR number
+        uses: actions/upload-artifact@v4
+        with:
+          name: pr-number
+          path: pr-number.txt
+          retention-days: 1

.github/workflows/claude.yml

@@ -0,0 +1,50 @@
+name: Claude Code
+
+on:
+  issue_comment:
+    types: [created]
+  pull_request_review_comment:
+    types: [created]
+  issues:
+    types: [opened, assigned]
+  pull_request_review:
+    types: [submitted]
+
+jobs:
+  claude:
+    if: |
+      (github.event_name == 'issue_comment' && contains(github.event.comment.body, '@claude')) ||
+      (github.event_name == 'pull_request_review_comment' && contains(github.event.comment.body, '@claude')) ||
+      (github.event_name == 'pull_request_review' && contains(github.event.review.body, '@claude')) ||
+      (github.event_name == 'issues' && (contains(github.event.issue.body, '@claude') || contains(github.event.issue.title, '@claude')))
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      pull-requests: read
+      issues: read
+      id-token: write
+      actions: read # Required for Claude to read CI results on PRs
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 1
+
+      - name: Run Claude Code
+        id: claude
+        uses: anthropics/claude-code-action@v1
+        with:
+          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
+
+          # This is an optional setting that allows Claude to read CI results on PRs
+          additional_permissions: |
+            actions: read
+
+          # Optional: Give a custom prompt to Claude. If this is not specified, Claude will perform the instructions specified in the comment that tagged it.
+          # prompt: 'Update the pull request description to include a summary of changes.'
+
+          # Optional: Add claude_args to customize behavior and configuration
+          # See https://github.com/anthropics/claude-code-action/blob/main/docs/usage.md
+          # or https://code.claude.com/docs/en/cli-reference for available options
+          # claude_args: '--allowed-tools Bash(gh pr *)'
+

.gitignore

@@ -18,6 +18,7 @@ testdata/**/*
 !assets/js/**/*.yaml
 !package*.json
 !evcc.dist.yaml
+!.coderabbit.yaml
 !tests/**/*.evcc.yaml
 !tests/**/*.tpl.yaml
 !tsconfig.json
@@ -49,3 +50,5 @@ asset-stats.html
 /evcc.db
 /tsconfig.tsbuildinfo
 *storybook.log
+/graphify-out/
+.graphifyignore

AGENTS.md

@@ -90,8 +90,10 @@ Deep documentation on specific subsystems is available in `docs/agents/`. Load w
 
 - No em dashes (—) in comments, commit messages, or docs. Use periods, commas, or colons
 - Project name is `evcc`, always lowercase
+- In user-facing strings, only mention `evcc` when needed to understand the context. Inside evcc's own UI the self-reference is usually redundant
 - Acronyms uppercase in prose: OCPP, MQTT, HEMS, SoC
 - Terminology: German "Phasensaldierung" (meter netting signed power across phases each instant) is "summative energy measurement" in English. Avoid "phase balancing" (means load balancing) and "net metering" (a billing scheme)
+- Terminology: the top-level load management circuit is "root circuit" in English, "Hauptstromkreis" in German
 - Commit subjects: `Component: short description`, no trailing period. Sub-scope in parens: `Meter (Home Assistant): ...`. Use `chore:`/`fix:`/`docs:` only for non-feature changes
 
 ## Comment Style
@@ -175,6 +177,7 @@ Deep documentation on specific subsystems is available in `docs/agents/`. Load w
 
 ### State Management
 
+- Never access the store from sub-components; keep them stateless and pass the values they need as props (emit events back to the parent). Only top-level views read from the store. This keeps components reusable and testable (e.g. Storybook should never mock the store).
 - Use `reactive()` from Vue for simple global state
 - Implement property setters for nested object updates using helper functions
 - Use localStorage with reactive wrappers for persistent settings
@@ -255,7 +258,9 @@ Deep documentation on specific subsystems is available in `docs/agents/`. Load w
 - Use `expectModalVisible()` and `expectModalHidden()` helpers
 - Test configuration persistence across application restarts
 - Standard structure: import `{ start, stop, baseUrl }` from `./evcc`, use `test.afterEach(stop)`
-- Never use fixed timeouts, use existance of elements or wait for network idle
+- Never use fixed timeouts. Wait on element state (visibility, count, value) instead.
+- Never use `page.waitForLoadState("networkidle")`. SPAs keep emitting requests (websockets, polling), so it either races or hangs. Wait for the specific element / value you need instead.
+- Keep test names and describe titles short and concrete. They should complement each other, not repeat. Prefer `describe("aux meter") test("create")` over `describe("aux meter") test("create aux meter and verify it appears")`. Drop scenario filler like "and lands in section", "appears correctly", "ensure".
 
 ## Device Integration & Configuration
 
@@ -264,6 +269,7 @@ Deep documentation on specific subsystems is available in `docs/agents/`. Load w
 - Device types: chargers, meters, vehicles, tariffs
 - Plugin protocols: Modbus, HTTP, MQTT, JavaScript, Go
 - Define device capabilities and configuration in templates at `templates/definition/[type]/`
+- Don't restate param properties that `util/templates/defaults.yaml` already defines for that param name. Properties (description, help, type, unit, default, example, required, advanced, mask, private, usages, …) are inherited from defaults; only specify a property in a template to give it a *different* value. Restating the same value is redundant duplication: reference the param by `name` alone.
 - Test templates: `evcc --template-type [type] --template [file]`
 - Update docs after template changes: `make docs`
 - When implementing or debugging against a third-party device library (eebus-go/ship/spine-go, ocpp-go, modbus/SunSpec), consult the library's current upstream documentation before coding rather than relying on recalled API details
@@ -307,3 +313,5 @@ Structure PR descriptions in this order. No headlines. Be concise.
    ```
 
 Avoid file paths, line numbers, or code listings reproduced from the diff. Include a code snippet only when it conveys the contract (event shape, API signature) more clearly than prose. No testing checklists, no co-author footers, no generator footers.
+
+Never state that `go build`, `go vet`, `go test -race`, or `gofmt` pass (or any "all checks/tests green" phrasing). These are non-negotiable givens that must already be fulfilled, not noteworthy results.

api/feature.go

@@ -4,17 +4,18 @@ type Feature int
 
 //go:generate go tool enumer -type Feature -text
 const (
-	_                Feature = iota
-	CoarseCurrent            // charger
-	IntegratedDevice         // charger - always connected - no vehicle, no charging sessions
-	SwitchDevice             // charger - no current control - heat pumps or switch sockets
-	Heating                  // charger - heating device - soc ist temperature (°C)
-	Continuous               // charger - heating device where disabled means "normal operation"
-	Average                  // tariff
-	Cacheable                // tariff
-	Offline                  // vehicle
-	Retryable                // vehicle
-	Streaming                // vehicle
-	WelcomeCharge            // vehicle
-	ClimaterDisabled         // vehicle - ignore climater state for charge control
+	_                  Feature = iota
+	CoarseCurrent              // charger
+	IntegratedDevice           // charger - always connected - no vehicle, no charging sessions
+	SwitchDevice               // charger - no current control - heat pumps or switch sockets
+	Heating                    // charger - heating device - soc ist temperature (°C)
+	Continuous                 // charger - heating device where disabled means "normal operation"
+	Average                    // tariff
+	Cacheable                  // tariff
+	Offline                    // vehicle
+	Retryable                  // vehicle
+	Streaming                  // vehicle
+	WelcomeCharge              // vehicle
+	ClimaterDisabled           // vehicle - ignore climater state for charge control
+	AutodetectDisabled         // vehicle - do not try to identify vehicle by status
 )

api/feature_enumer.go

@@ -54,6 +54,17 @@
 	--evcc-accent2: var(--evcc-darker-green);
 	--evcc-accent3: var(--evcc-darkest-green);
 
+	/* heating temperature bar: thermometer scale (cold -> hot) */
+	--evcc-temp-cold: #1b6fba;
+	--evcc-temp-warm: #ffa700;
+	--evcc-temp-hot: #e30022;
+	--evcc-heating-gradient: linear-gradient(
+		90deg,
+		var(--evcc-temp-cold) 5%,
+		var(--evcc-temp-warm) 20%,
+		var(--evcc-temp-warm) 40%,
+		var(--evcc-temp-hot) 95%
+	);
 	--evcc-transition-slow: 1000ms;
 	--evcc-transition-medium: 500ms;
 	--evcc-transition-fast: 250ms;