fix: address remaining shell injection and model name issues

- Move github.event.label.name to env var in Remove QA label step
- Move github.head_ref to env var RAW_BRANCH in Deploy videos step
- Update gpt-5.4 default model to gpt-4o in qa-video-review.ts

Co-authored-by: snomiao <7323030+snomiao@users.noreply.github.com>

.claude/skills/comfy-qa/SKILL.md

@@ -0,0 +1,358 @@
+---
+name: comfy-qa
+description: 'Comprehensive QA of ComfyUI frontend. Navigates all routes, tests all interactive features using Playwright MCP, generates a report, and submits a draft PR. Works in CI and local environments, cross-platform.'
+---
+
+# ComfyUI Frontend QA Skill
+
+Perform comprehensive quality assurance of the ComfyUI frontend application by navigating all routes, clicking interactive elements, and testing features. Generate a structured report and submit it as a draft PR.
+
+## Prerequisites
+
+- Node.js 18+
+- `pnpm` package manager
+- `gh` CLI (authenticated)
+- Playwright MCP server (browser automation)
+
+## Step 1: Environment Detection & Setup
+
+Detect the runtime environment and ensure the app is accessible.
+
+### CI Environment
+
+If `CI=true` is set:
+
+1. The ComfyUI backend is pre-configured in the CI container (`ghcr.io/comfy-org/comfyui-ci-container`)
+2. Frontend dist is already built and served by the backend
+3. Server runs at `http://127.0.0.1:8188`
+4. Skip user prompts — run fully automated
+
+### Local Environment
+
+If `CI` is not set:
+
+1. **Ask the user**: "Is a ComfyUI server already running? If so, what URL? (default: http://127.0.0.1:8188)"
+   - If yes: use the provided URL
+   - If no: offer to start one:
+
+     ```bash
+     # Option A: Use existing ComfyUI installation
+     # Ask for the path to ComfyUI, then:
+     cd <comfyui_path>
+     python main.py --cpu --multi-user --front-end-root <frontend_dist_path> &
+
+     # Option B: Build frontend and use preview server (no backend features)
+     pnpm build && pnpm preview &
+     ```
+
+2. Wait for server readiness by polling the URL (retry with 2s intervals, 60s timeout)
+
+### Browser Automation Setup
+
+Use **Playwright MCP** for browser interaction. The agent should use whichever Playwright MCP is available in its tool set:
+
+- **Claude Code**: Uses `remote-chrome` MCP or built-in browser tools
+- **Interactive mode**: User can opt for `npx @anthropic-ai/mcp-server-playwright` or `npx @playwright/mcp@latest --port 3100` for headless
+- **User's real browser**: `npx @playwright/mcp@latest --extension` — lets QA happen in the user's actual browser session
+
+Determine the server URL. For local dev servers behind proxies, adjust the URL accordingly (e.g., `https://[port].stukivx.xyz` pattern if configured).
+
+## Step 2: QA Test Plan
+
+Navigate to the application URL and systematically test each area below. For each test, record:
+
+- **Status**: pass / fail / skip (with reason)
+- **Notes**: any issues, unexpected behavior, or visual glitches
+- **Screenshots**: take screenshots of failures or notable states
+
+### 2.1 Application Load & Routes
+
+| Test              | Steps                                                        |
+| ----------------- | ------------------------------------------------------------ |
+| Root route loads  | Navigate to `/` — GraphView should render with canvas        |
+| User select route | Navigate to `/user-select` — user selection UI should appear |
+| Default redirect  | If multi-user mode, `/` redirects to `/user-select` first    |
+| 404 handling      | Navigate to `/nonexistent` — should handle gracefully        |
+
+### 2.2 Canvas & Graph View
+
+| Test                      | Steps                                                          |
+| ------------------------- | -------------------------------------------------------------- |
+| Canvas renders            | The LiteGraph canvas is visible and interactive                |
+| Pan canvas                | Click and drag on empty canvas area                            |
+| Zoom in/out               | Use scroll wheel or Alt+=/Alt+-                                |
+| Fit view                  | Press `.` key — canvas fits to content                         |
+| Add node via double-click | Double-click canvas to open search, type "KSampler", select it |
+| Add node via search       | Open search box, find and add a node                           |
+| Delete node               | Select a node, press Delete key                                |
+| Connect nodes             | Drag from output slot to input slot                            |
+| Disconnect nodes          | Right-click a link and remove, or drag from connected slot     |
+| Multi-select              | Shift+click or drag-select multiple nodes                      |
+| Copy/Paste                | Select nodes, Ctrl+C then Ctrl+V                               |
+| Undo/Redo                 | Make changes, Ctrl+Z to undo, Ctrl+Y to redo                   |
+| Node context menu         | Right-click a node — menu appears with all expected options    |
+| Canvas context menu       | Right-click empty canvas — menu appears                        |
+
+### 2.3 Node Operations
+
+| Test                | Steps                                                      |
+| ------------------- | ---------------------------------------------------------- |
+| Bypass node         | Select node, Ctrl+B — node shows bypass state              |
+| Mute node           | Select node, Ctrl+M — node shows muted state               |
+| Collapse node       | Select node, Alt+C — node collapses                        |
+| Pin node            | Select node, press P — node becomes pinned                 |
+| Rename node         | Double-click node title — edit mode activates              |
+| Node color          | Right-click > Color — color picker works                   |
+| Group nodes         | Select multiple nodes, Ctrl+G — group created              |
+| Ungroup             | Right-click group > Ungroup                                |
+| Widget interactions | Toggle checkboxes, adjust sliders, type in text fields     |
+| Combo widget        | Click dropdown widgets — options appear and are selectable |
+
+### 2.4 Sidebar Tabs
+
+| Test                   | Steps                                                  |
+| ---------------------- | ------------------------------------------------------ |
+| Workflows tab          | Press W — workflows sidebar opens with saved workflows |
+| Node Library tab       | Press N — node library opens with categories           |
+| Model Library tab      | Press M — model library opens                          |
+| Assets tab             | Press A — assets browser opens                         |
+| Tab toggle             | Press same key again — sidebar closes                  |
+| Search in sidebar      | Type in search box — results filter                    |
+| Drag node from library | Drag a node from library onto canvas                   |
+
+### 2.5 Topbar & Workflow Tabs
+
+| Test                 | Steps                                                  |
+| -------------------- | ------------------------------------------------------ |
+| Workflow tab display | Current workflow name shown in tab bar                 |
+| New workflow         | Ctrl+N — new blank workflow created                    |
+| Rename workflow      | Double-click workflow tab                              |
+| Tab context menu     | Right-click workflow tab — menu with Close/Rename/etc. |
+| Multiple tabs        | Open multiple workflows, switch between them           |
+| Queue button         | Click Queue/Run button — prompt queues                 |
+| Batch count          | Click batch count editor, change value                 |
+| Menu hamburger       | Click hamburger menu — options appear                  |
+
+### 2.6 Settings Dialog
+
+| Test             | Steps                                                |
+| ---------------- | ---------------------------------------------------- |
+| Open settings    | Press Ctrl+, or click settings button                |
+| Settings tabs    | Navigate through all setting categories              |
+| Change a setting | Toggle a boolean setting — it persists after closing |
+| Search settings  | Type in settings search box — results filter         |
+| Keybindings tab  | Navigate to keybindings panel                        |
+| About tab        | Navigate to about panel — version info shown         |
+| Close settings   | Press Escape or click close button                   |
+
+### 2.7 Bottom Panel
+
+| Test                | Steps                                  |
+| ------------------- | -------------------------------------- |
+| Toggle panel        | Press Ctrl+` — bottom panel opens      |
+| Logs tab            | Logs/terminal tab shows server output  |
+| Shortcuts tab       | Shortcuts reference is displayed       |
+| Keybindings display | Press Ctrl+Shift+K — keybindings panel |
+
+### 2.8 Execution & Queue
+
+| Test           | Steps                                                 |
+| -------------- | ----------------------------------------------------- |
+| Queue prompt   | Load default workflow, click Queue — execution starts |
+| Queue progress | Progress indicator shows during execution             |
+| Interrupt      | Press Ctrl+Alt+Enter during execution — interrupts    |
+| Job history    | Open job history sidebar — past executions listed     |
+| Clear history  | Clear execution history via menu                      |
+
+### 2.9 Workflow File Operations
+
+| Test            | Steps                                             |
+| --------------- | ------------------------------------------------- |
+| Save workflow   | Ctrl+S — workflow saves (check for prompt if new) |
+| Open workflow   | Ctrl+O — file picker or workflow browser opens    |
+| Export JSON     | Menu > Export — workflow JSON downloads           |
+| Import workflow | Drag a .json workflow file onto canvas            |
+| Load default    | Menu > Load Default — default workflow loads      |
+| Clear workflow  | Menu > Clear — canvas clears (after confirmation) |
+
+### 2.10 Advanced Features
+
+| Test            | Steps                                             |
+| --------------- | ------------------------------------------------- |
+| Minimap         | Alt+M — minimap toggle                            |
+| Focus mode      | Toggle focus mode                                 |
+| Canvas lock     | Press H to lock, V to unlock                      |
+| Link visibility | Ctrl+Shift+L — toggle links                       |
+| Subgraph        | Select nodes > Ctrl+Shift+E — convert to subgraph |
+
+### 2.11 Error Handling
+
+| Test                  | Steps                                        |
+| --------------------- | -------------------------------------------- |
+| Missing nodes dialog  | Load workflow with non-existent node types   |
+| Missing models dialog | Trigger missing model warning                |
+| Network error         | Disconnect backend, verify graceful handling |
+| Invalid workflow      | Try loading malformed JSON                   |
+
+### 2.12 Responsive & Accessibility
+
+| Test                | Steps                                 |
+| ------------------- | ------------------------------------- |
+| Window resize       | Resize browser window — layout adapts |
+| Keyboard navigation | Tab through interactive elements      |
+| Sidebar resize      | Drag sidebar edge to resize           |
+
+## Step 3: Generate Report
+
+After completing all tests, generate a markdown report file.
+
+### Report Location
+
+```
+docs/qa/YYYY-MM-DD-NNN-report.md
+```
+
+Where:
+
+- `YYYY-MM-DD` is today's date
+- `NNN` is a zero-padded increment index (001, 002, etc.)
+
+To determine the increment, check existing files:
+
+```bash
+ls docs/qa/ | grep "$(date +%Y-%m-%d)" | wc -l
+```
+
+### Report Template
+
+```markdown
+# QA Report: ComfyUI Frontend
+
+**Date**: YYYY-MM-DD
+**Environment**: CI / Local (OS, Browser)
+**Frontend Version**: (git sha or version)
+**Agent**: Claude / Codex / Other
+**Server URL**: http://...
+
+## Summary
+
+| Category        | Pass | Fail | Skip | Total |
+| --------------- | ---- | ---- | ---- | ----- |
+| Routes & Load   |      |      |      |       |
+| Canvas          |      |      |      |       |
+| Node Operations |      |      |      |       |
+| Sidebar         |      |      |      |       |
+| Topbar          |      |      |      |       |
+| Settings        |      |      |      |       |
+| Bottom Panel    |      |      |      |       |
+| Execution       |      |      |      |       |
+| File Operations |      |      |      |       |
+| Advanced        |      |      |      |       |
+| Error Handling  |      |      |      |       |
+| Responsive      |      |      |      |       |
+| **Total**       |      |      |      |       |
+
+## Results
+
+### Routes & Load
+
+- [x] Root route loads — pass
+- [ ] ...
+
+### Canvas & Graph View
+
+- [x] Canvas renders — pass
+- [ ] ...
+
+(repeat for each category)
+
+## Issues Found
+
+### Issue 1: [Title]
+
+- **Severity**: critical / major / minor / cosmetic
+- **Steps to reproduce**: ...
+- **Expected**: ...
+- **Actual**: ...
+- **Screenshot**: (if available)
+
+## Notes
+
+Any additional observations, performance notes, or suggestions.
+```
+
+## Step 4: Commit and Push Report
+
+### In CI (when `CI=true`)
+
+Commit and push the report directly to the current branch. Do **not** create
+a new PR — the CI workflow handles posting results as a PR comment.
+
+```bash
+git add docs/qa/
+git commit -m "docs: add QA report $(date +%Y-%m-%d)"
+git push
+```
+
+### Local / interactive use
+
+When running locally, create a draft PR after committing:
+
+```bash
+# Ensure on a feature branch
+BRANCH_NAME="qa/$(date +%Y-%m-%d)-$(git rev-parse --short HEAD)"
+git checkout -b "$BRANCH_NAME" 2>/dev/null || git checkout "$BRANCH_NAME"
+
+git add docs/qa/
+git commit -m "docs: add QA report $(date +%Y-%m-%d)
+
+Automated QA report covering all frontend routes and features."
+git push -u origin "$BRANCH_NAME"
+
+# Create draft PR assigned to comfy-pr-bot
+gh pr create \
+  --draft \
+  --title "QA Report: $(date +%Y-%m-%d)" \
+  --body "## QA Report
+
+Automated frontend QA run covering all routes and interactive features.
+
+See \`docs/qa/\` for the full report.
+
+/cc @comfy-pr-bot" \
+  --assignee comfy-pr-bot
+```
+
+## Execution Notes
+
+### Cross-Platform Considerations
+
+- **Windows**: Use `pwsh` or `cmd` equivalents for shell commands. `gh` CLI works on all platforms.
+- **macOS**: Keyboard shortcuts use Cmd instead of Ctrl in the actual app, but Playwright sends OS-appropriate keys.
+- **Linux**: Primary CI platform. Screenshot baselines are Linux-only.
+
+### Agent Compatibility
+
+This skill is written to be agent-agnostic:
+
+- **Claude Code**: Has native Playwright MCP via `remote-chrome` or configured MCP servers
+- **Codex**: Can use `npx @playwright/mcp@latest` for browser automation
+- **Other agents**: Any agent with browser automation tools can follow these instructions
+
+The key requirement is the ability to:
+
+1. Navigate to URLs
+2. Take snapshots/screenshots
+3. Click elements
+4. Type text
+5. Press keyboard shortcuts
+6. Read page content
+
+### Tips for Reliable QA
+
+1. **Wait for page stability** before interacting — check that elements are visible and enabled
+2. **Take a snapshot after each major navigation** to verify state
+3. **Don't use fixed timeouts** — poll for expected conditions
+4. **Record the full page snapshot** at the start for baseline comparison
+5. **If a test fails**, document it and continue — don't abort the entire QA run
+6. **Group related tests** — complete one category before moving to the next

.github/actions/setup-comfyui-server/action.yaml

@@ -44,7 +44,6 @@ runs:
         python -m pip install --upgrade pip
         pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
         pip install -r requirements.txt
-        pip install wait-for-it
 
     - name: Start ComfyUI server
       if: ${{ inputs.launch_server == 'true' }}
@@ -52,4 +51,8 @@ runs:
       working-directory: ComfyUI
       run: |
         python main.py --cpu --multi-user --front-end-root ../dist ${{ inputs.extra_server_params }} &
-        wait-for-it --service 127.0.0.1:8188 -t 600
+        for i in $(seq 1 300); do
+          curl -sf http://127.0.0.1:8188/api/system_stats >/dev/null 2>&1 && echo "Server ready" && exit 0
+          sleep 2
+        done
+        echo "::error::ComfyUI server did not start within 600s" && exit 1

.github/workflows/pr-qa.yaml

@@ -0,0 +1,564 @@
+# Automated QA of ComfyUI frontend using Claude CLI + Playwright MCP.
+# Two modes:
+#   Focused (qa-changes label): Linux-only, tests areas affected by PR changes
+#   Full (qa-full label): 3-OS matrix, full test plan
+name: 'PR: QA'
+
+on:
+  pull_request:
+    types: [labeled]
+    branches: [main]
+  workflow_dispatch:
+    inputs:
+      mode:
+        description: 'QA mode'
+        type: choice
+        options: [focused, full]
+        default: focused
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  resolve-matrix:
+    runs-on: ubuntu-latest
+    outputs:
+      os: ${{ steps.set.outputs.os }}
+      mode: ${{ steps.set.outputs.mode }}
+      skip: ${{ steps.set.outputs.skip }}
+    steps:
+      - name: Determine QA mode
+        id: set
+        env:
+          LABEL: ${{ github.event.label.name }}
+          EVENT_ACTION: ${{ github.event.action }}
+          EVENT_NAME: ${{ github.event_name }}
+          INPUT_MODE: ${{ inputs.mode }}
+        run: |
+          FULL=false
+
+          # Only run on label events if it's one of our labels
+          if [ "$EVENT_ACTION" = "labeled" ] && \
+             [ "$LABEL" != "qa-changes" ] && [ "$LABEL" != "qa-full" ]; then
+            echo "skip=true" >> "$GITHUB_OUTPUT"
+          fi
+
+          # Full QA triggers
+          if [ "$EVENT_NAME" = "workflow_dispatch" ] && \
+             [ "$INPUT_MODE" = "full" ]; then
+            FULL=true
+          fi
+          if [ "$LABEL" = "qa-full" ]; then
+            FULL=true
+          fi
+
+          if [ "$FULL" = "true" ]; then
+            echo 'os=["ubuntu-latest","macos-latest","windows-latest"]' >> "$GITHUB_OUTPUT"
+            echo "mode=full" >> "$GITHUB_OUTPUT"
+          else
+            echo 'os=["ubuntu-latest"]' >> "$GITHUB_OUTPUT"
+            echo "mode=focused" >> "$GITHUB_OUTPUT"
+          fi
+          echo "Mode: $([ "$FULL" = "true" ] && echo full || echo focused)"
+
+  qa:
+    needs: resolve-matrix
+    if: needs.resolve-matrix.outputs.skip != 'true'
+    strategy:
+      fail-fast: false
+      matrix:
+        os: ${{ fromJson(needs.resolve-matrix.outputs.os) }}
+    runs-on: ${{ matrix.os }}
+    timeout-minutes: 60
+    permissions:
+      contents: write
+      pull-requests: write
+    env:
+      QA_MODE: ${{ needs.resolve-matrix.outputs.mode }}
+    steps:
+      - name: Set QA artifacts path
+        shell: bash
+        run: echo "QA_ARTIFACTS=$RUNNER_TEMP/qa-artifacts" >> "$GITHUB_ENV"
+
+      - name: Checkout repository
+        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
+        with:
+          fetch-depth: 0
+          ref: ${{ github.head_ref || github.ref }}
+          token: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Setup frontend
+        uses: ./.github/actions/setup-frontend
+        with:
+          include_build_step: true
+
+      - name: Setup and start ComfyUI server
+        uses: ./.github/actions/setup-comfyui-server
+        with:
+          launch_server: 'true'
+
+      - name: Wait for ComfyUI server
+        shell: bash
+        run: |
+          echo "Waiting for ComfyUI server..."
+          for i in $(seq 1 60); do
+            if curl -sf http://127.0.0.1:8188/api/system_stats >/dev/null 2>&1; then
+              echo "Server ready"; exit 0
+            fi; sleep 2
+          done
+          echo "::error::Server timeout"; exit 1
+
+      - name: Install Playwright for MCP
+        shell: bash
+        run: pnpm exec playwright install chromium --with-deps
+
+      - name: Install Claude Code
+        shell: bash
+        run: npm install -g @anthropic-ai/claude-code@2.1.71
+
+      # --- Virtual display + recording: Linux ---
+      - name: Setup Xvfb (Linux)
+        if: runner.os == 'Linux'
+        run: |
+          sudo apt-get update -qq && sudo apt-get install -y -qq xvfb ffmpeg x11-xserver-utils >/dev/null 2>&1
+          Xvfb :99 -screen 0 1280x720x24 -ac -nolisten tcp &
+          echo $! > "${{ runner.temp }}/xvfb.pid"
+          sleep 2
+          echo "DISPLAY=:99" >> "$GITHUB_ENV"
+          # Set black background and normal cursor (removes the default X cursor)
+          DISPLAY=:99 xsetroot -solid black -cursor_name left_ptr 2>/dev/null || true
+
+      - name: Start recording (Linux)
+        if: runner.os == 'Linux'
+        continue-on-error: true
+        run: |
+          mkdir -p "$QA_ARTIFACTS"
+          ffmpeg -y -f x11grab -video_size 1280x720 -framerate 10 \
+            -i :99.0 -c:v libx264 -preset ultrafast -crf 28 \
+            -pix_fmt yuv420p "$QA_ARTIFACTS/qa-session.mp4" &
+          echo $! > "${{ runner.temp }}/ffmpeg.pid"
+          sleep 1
+
+      # --- Recording: macOS (avfoundation) ---
+      - name: Install ffmpeg (macOS)
+        if: runner.os == 'macOS'
+        run: brew install ffmpeg
+
+      - name: Grant screen recording permission (macOS)
+        if: runner.os == 'macOS'
+        run: |
+          # Grant kTCCServiceScreenCapture to ffmpeg and bash so avfoundation
+          # recording works without the blocking permission dialog.
+          FFMPEG_PATH=$(which ffmpeg)
+          echo "Granting screen recording permission to: $FFMPEG_PATH"
+          # macOS 14+ (Sonoma/Sequoia) TCC.db schema has 17 columns
+          for BIN in "$FFMPEG_PATH" "/bin/bash"; do
+            sudo sqlite3 "/Library/Application Support/com.apple.TCC/TCC.db" \
+              "INSERT OR REPLACE INTO access VALUES('kTCCServiceScreenCapture','${BIN}',1,2,0,1,NULL,NULL,0,'UNUSED',NULL,0,$(date +%s),NULL,NULL,NULL,NULL);" 2>/dev/null \
+            || sudo sqlite3 "/Library/Application Support/com.apple.TCC/TCC.db" \
+              "INSERT OR REPLACE INTO access VALUES('kTCCServiceScreenCapture','${BIN}',1,2,0,1,NULL,NULL,NULL,'UNUSED',NULL,0,$(date +%s));" 2>/dev/null \
+            || echo "Warning: Could not grant TCC permission to ${BIN}"
+          done
+
+      - name: Start recording (macOS)
+        if: runner.os == 'macOS'
+        continue-on-error: true
+        run: |
+          mkdir -p "$QA_ARTIFACTS"
+          ffmpeg -f avfoundation -list_devices true -i "" 2>&1 || true
+          ffmpeg -y -f avfoundation -framerate 10 -capture_cursor 1 \
+            -i "Capture screen 0:none" -c:v libx264 -preset ultrafast -crf 28 \
+            -pix_fmt yuv420p "$QA_ARTIFACTS/qa-session.mp4" &
+          echo $! > "${{ runner.temp }}/ffmpeg.pid"
+          sleep 2
+
+      # --- Recording: Windows (gdigrab) ---
+      - name: Install ffmpeg (Windows)
+        if: runner.os == 'Windows'
+        shell: pwsh
+        run: choco install ffmpeg -y --no-progress
+
+      - name: Start recording (Windows)
+        if: runner.os == 'Windows'
+        continue-on-error: true
+        shell: bash
+        run: |
+          mkdir -p "$QA_ARTIFACTS"
+          ffmpeg -y -f gdigrab -framerate 10 -i desktop \
+            -c:v libx264 -preset ultrafast -crf 28 \
+            -pix_fmt yuv420p "$QA_ARTIFACTS/qa-session.mp4" &
+          echo $! > "${{ runner.temp }}/ffmpeg.pid"
+          sleep 1
+
+      - name: Create MCP config
+        shell: bash
+        run: |
+          cat > "${{ runner.temp }}/mcp-config.json" <<EOF
+          {"mcpServers":{"playwright":{"command":"npx","args":["@playwright/mcp@0.0.68"]}}}
+          EOF
+
+      - name: Get PR diff for focused QA
+        if: needs.resolve-matrix.outputs.mode == 'focused'
+        shell: bash
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          gh pr diff ${{ github.event.pull_request.number || '' }} \
+            --repo ${{ github.repository }} > "${{ runner.temp }}/pr-diff.txt" 2>/dev/null || \
+          git diff origin/main...HEAD > "${{ runner.temp }}/pr-diff.txt"
+
+          # Summarize changed files for the prompt
+          echo "Changed files:"
+          grep '^diff --git' "${{ runner.temp }}/pr-diff.txt" | \
+            sed 's|diff --git a/||;s| b/.*||' | sort -u | tee "${{ runner.temp }}/changed-files.txt"
+
+      - name: Write QA prompt
+        shell: bash
+        env:
+          BRANCH: ${{ github.head_ref || github.ref_name }}
+          PR_NUM: ${{ github.event.pull_request.number || 'N/A' }}
+          SHA: ${{ github.sha }}
+        run: |
+          OS_LOWER=$(echo "$RUNNER_OS" | tr '[:upper:]' '[:lower:]')
+
+          if [ "$QA_MODE" = "full" ]; then
+            cat > "${{ runner.temp }}/qa-prompt.txt" <<PROMPT
+          You are running a FULL automated QA pass on the ComfyUI frontend.
+          Read the file .claude/skills/comfy-qa/SKILL.md and follow the FULL QA test plan.
+
+          Environment: CI=true, OS=${{ runner.os }}
+          Server URL: http://127.0.0.1:8188
+          Branch: ${BRANCH}
+          PR: #${PR_NUM}
+          Commit: ${SHA}
+
+          1. Use playwright MCP tools to navigate http://127.0.0.1:8188
+          2. Run the FULL QA test plan from the skill file
+          3. Take screenshots of failures or notable states
+          4. Save report to docs/qa/ as YYYY-MM-DD-NNN-${OS_LOWER}-report.md
+          5. Commit and push the report to this branch
+
+          Do NOT create a new PR. Do NOT post PR comments.
+          Skip tests not available in CI (file dialogs, GPU execution).
+          PROMPT
+          else
+            cat > "${{ runner.temp }}/qa-prompt.txt" <<PROMPT
+          You are running a FOCUSED QA pass on a pull request to the ComfyUI frontend.
+          Your goal is to verify that the changes in this PR work correctly and don't break related functionality.
+
+          Environment: CI=true, OS=${{ runner.os }}
+          Server URL: http://127.0.0.1:8188
+          Branch: ${BRANCH}
+          PR: #${PR_NUM}
+          Commit: ${SHA}
+
+          CHANGED FILES:
+          $(cat "${{ runner.temp }}/changed-files.txt" 2>/dev/null || echo "Unknown")
+
+          DIFF (truncated to 500 lines):
+          $(head -500 "${{ runner.temp }}/pr-diff.txt" 2>/dev/null || echo "No diff available")
+
+          Instructions:
+          1. Read the diff above to understand what changed in this PR
+          2. Use playwright MCP tools to navigate http://127.0.0.1:8188
+          3. Test the specific UI areas affected by these changes
+          4. Also do a quick smoke test of core functionality (app loads, canvas renders, sidebar works)
+          5. Take screenshots of any failures or the areas you tested
+          6. Save a concise report to docs/qa/ as YYYY-MM-DD-NNN-${OS_LOWER}-report.md
+          7. Commit and push the report to this branch
+
+          Focus on:
+          - Does the changed functionality work as expected?
+          - Are there visual regressions in affected areas?
+          - Do related features still work?
+
+          Do NOT run the full QA test plan. Do NOT create a new PR. Do NOT post PR comments.
+          Skip tests not available in CI (file dialogs, GPU execution).
+          PROMPT
+          fi
+
+      - name: Run Claude QA
+        shell: bash
+        env:
+          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
+          CI: 'true'
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          MAX_TURNS=128
+          if [ "$QA_MODE" = "focused" ]; then MAX_TURNS=30; fi
+
+          cat "${{ runner.temp }}/qa-prompt.txt" | claude --print --verbose \
+            --max-turns "$MAX_TURNS" \
+            --mcp-config "${{ runner.temp }}/mcp-config.json" \
+            --allowedTools "mcp__playwright__browser_navigate,mcp__playwright__browser_snapshot,mcp__playwright__browser_click,mcp__playwright__browser_type,mcp__playwright__browser_press_key,mcp__playwright__browser_take_screenshot,mcp__playwright__browser_hover,mcp__playwright__browser_drag,mcp__playwright__browser_select_option,mcp__playwright__browser_handle_dialog,mcp__playwright__browser_tab_list,mcp__playwright__browser_tab_new,mcp__playwright__browser_tab_select,mcp__playwright__browser_tab_close,mcp__playwright__browser_console_messages,mcp__playwright__browser_resize,mcp__playwright__browser_wait_for,Bash(git add:*),Bash(git commit:*),Bash(git push:*),Bash(git status:*),Bash(git log:*),Bash(git diff:*),Bash(date:*),Bash(ls:*),Bash(mkdir:*),Read,Write,Edit,Glob,Grep"
+
+      - name: Stop recording
+        if: always()
+        shell: bash
+        run: |
+          PID_FILE="${{ runner.temp }}/ffmpeg.pid"
+          if [ -f "$PID_FILE" ]; then
+            if [ "$RUNNER_OS" = "Windows" ]; then
+              # Graceful stop: taskkill without /F sends WM_CLOSE so ffmpeg
+              # can finalize the mp4 (write moov atom). Force-kill as fallback.
+              taskkill //PID $(cat "$PID_FILE") 2>/dev/null || true
+              sleep 5
+              taskkill //F //PID $(cat "$PID_FILE") 2>/dev/null || true
+            else
+              kill -INT $(cat "$PID_FILE") 2>/dev/null || true
+              sleep 3; kill $(cat "$PID_FILE") 2>/dev/null || true
+            fi
+          fi
+          [ -f "$QA_ARTIFACTS/qa-session.mp4" ] && \
+            echo "Video: $(du -h "$QA_ARTIFACTS/qa-session.mp4" | cut -f1)" || \
+            echo "No video (non-fatal)"
+
+      - name: Collect artifacts
+        if: always()
+        shell: bash
+        run: |
+          mkdir -p "$QA_ARTIFACTS"
+          cp -r docs/qa/* "$QA_ARTIFACTS/" 2>/dev/null || true
+          ls -la "$QA_ARTIFACTS/" || true
+
+      - name: Upload QA artifacts
+        if: always()
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v6.2.0
+        with:
+          name: qa-report-${{ runner.os }}-${{ github.run_id }}
+          path: ${{ env.QA_ARTIFACTS }}/
+          retention-days: 14
+
+      - name: Cleanup (Linux)
+        if: always() && runner.os == 'Linux'
+        run: kill $(cat "${{ runner.temp }}/xvfb.pid") 2>/dev/null || true
+
+  report:
+    needs: [resolve-matrix, qa]
+    if: always() && github.event.pull_request.number && github.event.pull_request.head.repo.full_name == github.repository
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    steps:
+      - name: Download QA artifacts
+        uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131 # v7.0.0
+        with:
+          path: qa-artifacts
+          pattern: qa-report-*
+
+      - name: Install ffmpeg
+        run: sudo apt-get update -qq && sudo apt-get install -y -qq ffmpeg >/dev/null 2>&1
+
+      - name: Deploy videos to Cloudflare Pages
+        id: deploy-videos
+        env:
+          CLOUDFLARE_API_TOKEN: ${{ secrets.CLOUDFLARE_API_TOKEN }}
+          CLOUDFLARE_ACCOUNT_ID: ${{ secrets.CLOUDFLARE_ACCOUNT_ID }}
+          RAW_BRANCH: ${{ github.head_ref || github.ref_name }}
+        run: |
+          npm install -g wrangler@^4.0.0 >/dev/null 2>&1
+
+          DEPLOY_DIR=$(mktemp -d)
+          mkdir -p "$DEPLOY_DIR"
+
+          for os in Linux macOS Windows; do
+            VID="qa-artifacts/qa-report-${os}-${{ github.run_id }}/qa-session.mp4"
+            if [ -f "$VID" ]; then
+              cp "$VID" "$DEPLOY_DIR/qa-${os}.mp4"
+              echo "Found ${os} video ($(du -h "$VID" | cut -f1))"
+
+              # Generate GIF thumbnail: 8s starting at 10s, 480px wide, 8fps
+              ffmpeg -y -ss 10 -i "$VID" -t 8 \
+                -vf "fps=8,scale=480:-1:flags=lanczos,split[s0][s1];[s0]palettegen=max_colors=64[p];[s1][p]paletteuse=dither=bayer" \
+                -loop 0 "$DEPLOY_DIR/qa-${os}-thumb.gif" 2>/dev/null \
+                || ffmpeg -y -i "$VID" -t 8 \
+                  -vf "fps=8,scale=480:-1:flags=lanczos,split[s0][s1];[s0]palettegen=max_colors=64[p];[s1][p]paletteuse=dither=bayer" \
+                  -loop 0 "$DEPLOY_DIR/qa-${os}-thumb.gif" 2>/dev/null \
+                || echo "GIF generation failed for ${os} (non-fatal)"
+            fi
+          done
+
+          # Build video cards HTML
+          CARDS=""
+          ICONS_Linux="&#x1F427;" ICONS_macOS="&#x1F34E;" ICONS_Windows="&#x1FA9F;"
+          for os in Linux macOS Windows; do
+            eval "ICON=\$ICONS_${os}"
+            if [ -f "$DEPLOY_DIR/qa-${os}.mp4" ]; then
+              CARDS="${CARDS}<div class=card><video controls autoplay muted loop preload=metadata><source src=qa-${os}.mp4 type=video/mp4></video><div class=card-body><span class=platform><span class=icon>${ICON}</span> ${os}</span><a class=download href=qa-${os}.mp4 download>Download</a></div></div>"
+            else
+              CARDS="${CARDS}<div class=card><div class=empty-card>No recording available</div><div class=card-body><span class=platform><span class=icon>${ICON}</span> ${os}</span><span class='badge missing'>Missing</span></div></div>"
+            fi
+          done
+
+          cat > "$DEPLOY_DIR/index.html" <<INDEXEOF
+          <!DOCTYPE html><html lang=en><head><meta charset=utf-8><meta name=viewport content="width=device-width,initial-scale=1"><title>QA Session Recordings</title>
+          <style>
+          *{margin:0;padding:0;box-sizing:border-box}body{background:#0d1117;color:#e6edf3;font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Helvetica,Arial,sans-serif;min-height:100vh;padding:2rem 1rem}.container{max-width:1200px;margin:0 auto}header{display:flex;align-items:center;gap:.75rem;margin-bottom:2rem;padding-bottom:1rem;border-bottom:1px solid #30363d}h1{font-size:1.5rem;font-weight:600}.meta{color:#8b949e;font-size:.875rem;margin-top:.25rem}.grid{display:grid;grid-template-columns:repeat(auto-fill,minmax(340px,1fr));gap:1.25rem}.card{background:#161b22;border:1px solid #30363d;border-radius:.5rem;overflow:hidden;transition:border-color .15s}.card:hover{border-color:#58a6ff}.card video{width:100%;display:block;background:#010409;aspect-ratio:16/9;object-fit:contain}.card-body{padding:.75rem 1rem;display:flex;align-items:center;justify-content:space-between}.platform{display:flex;align-items:center;gap:.5rem;font-weight:500}.icon{font-size:1.25rem}.badge{font-size:.75rem;padding:.125rem .5rem;border-radius:999px;background:#1f6feb33;color:#58a6ff;border:1px solid #1f6feb55}.badge.missing{background:#da363333;color:#f85149;border-color:#da363355}.empty-card{display:flex;align-items:center;justify-content:center;min-height:200px;color:#484f58;font-size:.875rem}a.download{color:#58a6ff;text-decoration:none;font-size:.8125rem}a.download:hover{text-decoration:underline}
+          </style></head><body><div class=container>
+          <header><svg width=28 height=28 viewBox="0 0 24 24" fill=none stroke=#58a6ff stroke-width=2 stroke-linecap=round stroke-linejoin=round><polygon points="23 7 16 12 23 17 23 7"/><rect x=1 y=5 width=15 height=14 rx=2 ry=2/></svg><div><h1>QA Session Recordings</h1><div class=meta>ComfyUI Frontend &middot; Automated QA</div></div></header>
+          <div class=grid>${CARDS}</div>
+          </div></body></html>
+          INDEXEOF
+
+          # 404 page so Cloudflare Pages returns proper 404 for missing files
+          # (instead of SPA fallback serving index.html)
+          cat > "$DEPLOY_DIR/404.html" <<'ERROREOF'
+          <!DOCTYPE html><html><head><meta charset=utf-8><title>404</title>
+          <style>body{background:#0d1117;color:#8b949e;font-family:sans-serif;display:flex;align-items:center;justify-content:center;min-height:100vh;margin:0}div{text-align:center}h1{color:#f85149;font-size:3rem;margin-bottom:.5rem}p{font-size:1rem}</style>
+          </head><body><div><h1>404</h1><p>File not found. The QA recording may have failed or been cancelled.</p></div></body></html>
+          ERROREOF
+
+          # Sanitize branch name for Cloudflare Pages URL (same rules CF uses)
+          BRANCH=$(echo "$RAW_BRANCH" | sed 's/[^a-zA-Z0-9-]/-/g' | sed 's/--*/-/g' | sed 's/^-//;s/-$//' | cut -c1-28)
+          URL=$(wrangler pages deploy "$DEPLOY_DIR" \
+            --project-name="comfyui-qa-videos" \
+            --branch="$BRANCH" 2>&1 \
+            | grep -oE 'https://[a-zA-Z0-9.-]+\.pages\.dev\S*' | head -1)
+
+          echo "url=${URL:-https://${BRANCH}.comfyui-qa-videos.pages.dev}" >> "$GITHUB_OUTPUT"
+          echo "Deployed to: ${URL}"
+
+      - name: Post QA comment on PR
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          VIDEO_BASE: ${{ steps.deploy-videos.outputs.url }}
+          QA_MODE: ${{ needs.resolve-matrix.outputs.mode }}
+        run: |
+          RUN="https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}"
+          COMMENT_MARKER="<!-- QA_REPORT_COMMENT -->"
+
+          MODE_BADGE="🔍 Focused"
+          if [ "$QA_MODE" = "full" ]; then MODE_BADGE="🔬 Full (3-OS)"; fi
+
+          # Build video section with GIF thumbnails linking to full videos
+          VIDEO_SECTION=""
+          for os in Linux macOS Windows; do
+            GIF_URL="${VIDEO_BASE}/qa-${os}-thumb.gif"
+            VID_URL="${VIDEO_BASE}/qa-${os}.mp4"
+            if curl -sf --head "$VID_URL" >/dev/null 2>&1; then
+              if curl -sf --head "$GIF_URL" >/dev/null 2>&1; then
+                VIDEO_SECTION="${VIDEO_SECTION}[![${os} QA](${GIF_URL})](${VID_URL})"$'\n'
+              else
+                VIDEO_SECTION="${VIDEO_SECTION}[${os} video](${VID_URL})"$'\n'
+              fi
+            fi
+          done
+
+          BODY=$(cat <<EOF
+          ${COMMENT_MARKER}
+          ## QA ${MODE_BADGE}
+
+          ${VIDEO_SECTION}
+          **Run**: [${RUN}](${RUN}) · [Download artifacts](${RUN}#artifacts) · [All videos](${VIDEO_BASE})
+          EOF
+          )
+
+          EXISTING=$(gh api "repos/${{ github.repository }}/issues/${{ github.event.pull_request.number }}/comments" \
+            --jq ".[] | select(.body | contains(\"${COMMENT_MARKER}\")) | .id" | head -1)
+
+          if [ -n "$EXISTING" ]; then
+            gh api --method PATCH "repos/${{ github.repository }}/issues/comments/${EXISTING}" \
+              --field body="$BODY"
+          else
+            gh pr comment ${{ github.event.pull_request.number }} \
+              --repo ${{ github.repository }} --body "$BODY"
+          fi
+
+      - name: Remove QA label
+        if: >-
+          github.event.label.name == 'qa-changes' ||
+          github.event.label.name == 'qa-full'
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          LABEL: ${{ github.event.label.name }}
+        run: |
+          gh pr edit ${{ github.event.pull_request.number }} \
+            --repo ${{ github.repository }} --remove-label "$LABEL"
+
+  video-review:
+    needs: [qa, report]
+    if: always() && github.event.pull_request.number && github.event.pull_request.head.repo.full_name == github.repository
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
+
+      - name: Setup Node.js
+        uses: actions/setup-node@6044e13b5dc448c55e2357c09f80417699197238 # v6.2.0
+        with:
+          node-version: 24
+
+      - name: Install pnpm
+        uses: pnpm/action-setup@fc06bc1257f339d1d5d8b3a19a8cae5388b55320 # v4.4.0
+        with:
+          version: 10
+
+      - name: Install dependencies
+        run: pnpm install --frozen-lockfile
+
+      - name: Install ffmpeg
+        run: sudo apt-get update -qq && sudo apt-get install -y -qq ffmpeg >/dev/null 2>&1
+
+      - name: Download QA artifacts
+        uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131 # v7.0.0
+        with:
+          path: qa-artifacts
+          pattern: qa-report-*
+
+      - name: Run video review
+        env:
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+        run: |
+          mkdir -p video-reviews
+          for vid in qa-artifacts/qa-report-*/qa-session.mp4; do
+            [ -f "$vid" ] || continue
+            echo "::group::Reviewing $vid"
+            pnpm exec tsx scripts/qa-video-review.ts \
+              --artifacts-dir qa-artifacts \
+              --output-dir video-reviews \
+              --video-file "$vid" \
+              --model gpt-4o || true
+            echo "::endgroup::"
+          done
+
+      - name: Post video review comment
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          MARKER="<!-- QA_VIDEO_REVIEW_COMMENT -->"
+          RUN="https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}"
+
+          # Combine per-platform reports
+          REPORT=""
+          for f in video-reviews/*-qa-video-report.md; do
+            [ -f "$f" ] || continue
+            [ -n "$REPORT" ] && REPORT="${REPORT}
+          ---
+          "
+            REPORT="${REPORT}$(cat "$f")"
+          done
+
+          [ -z "$REPORT" ] && REPORT="No video reports were generated."
+
+          BODY="${MARKER}
+          ## QA Video Report
+
+          ${REPORT}
+
+          ---
+          **Run**: [${RUN}](${RUN})"
+
+          EXISTING=$(gh api "repos/${{ github.repository }}/issues/${{ github.event.pull_request.number }}/comments" \
+            --jq ".[] | select(.body | contains(\"${MARKER}\")) | .id" | head -1)
+
+          if [ -n "$EXISTING" ]; then
+            gh api --method PATCH "repos/${{ github.repository }}/issues/comments/${EXISTING}" \
+              --field body="$BODY"
+          else
+            gh pr comment ${{ github.event.pull_request.number }} \
+              --repo ${{ github.repository }} --body "$BODY"
+          fi

docs/qa/screenshots/.gitignore

@@ -0,0 +1,5 @@
+*.png
+*.jpg
+*.jpeg
+*.webp
+*.gif

knip.config.ts

@@ -29,10 +29,12 @@ const config: KnipConfig = {
     },
     'packages/ingest-types': {
       project: ['src/**/*.{js,ts}'],
-      entry: ['src/index.ts']
+      entry: ['src/index.ts', 'openapi-ts.config.ts'],
+      // openapi-ts binary is provided by @hey-api/openapi-ts devDependency
+      ignoreBinaries: ['openapi-ts']
     }
   },
-  ignoreBinaries: ['python3', 'gh', 'generate'],
+  ignoreBinaries: ['python3', 'gh'],
   ignoreDependencies: [
     // Weird importmap things
     '@iconify-json/lucide',
@@ -40,9 +42,7 @@ const config: KnipConfig = {
     '@primeuix/forms',
     '@primeuix/styled',
     '@primeuix/utils',
-    '@primevue/icons',
-    // Used by lucideStrokePlugin.js (CSS @plugin)
-    '@iconify/utils'
+    '@primevue/icons'
   ],
   ignore: [
     // Auto generated API types
@@ -50,7 +50,6 @@ const config: KnipConfig = {
     'packages/registry-types/src/comfyRegistryTypes.ts',
     'packages/ingest-types/src/types.gen.ts',
     'packages/ingest-types/src/zod.gen.ts',
-    'packages/ingest-types/openapi-ts.config.ts',
     // Used by a custom node (that should move off of this)
     'src/scripts/ui/components/splitButton.ts',
     // Used by stacked PR (feat/glsl-live-preview)

package.json

@@ -39,6 +39,7 @@
     "preinstall": "pnpm dlx only-allow pnpm",
     "prepare": "husky || true && git config blame.ignoreRevsFile .git-blame-ignore-revs || true",
     "preview": "nx preview",
+    "qa:video-review": "tsx scripts/qa-video-review.ts",
     "storybook": "nx storybook",
     "storybook:desktop": "nx run @comfyorg/desktop-ui:storybook",
     "stylelint:fix": "stylelint --cache --fix '{apps,packages,src}/**/*.{css,vue}'",

scripts/qa-video-review.test.ts

@@ -0,0 +1,183 @@
+import { describe, expect, it } from 'vitest'
+
+import {
+  extractOutputText,
+  extractPlatformFromArtifactDirName,
+  pickLatestVideosByPlatform,
+  selectVideoCandidateByFile
+} from './qa-video-review'
+
+describe('extractPlatformFromArtifactDirName', () => {
+  it('extracts and normalizes known qa artifact directory names', () => {
+    expect(
+      extractPlatformFromArtifactDirName('qa-report-Windows-22818315023')
+    ).toBe('windows')
+    expect(
+      extractPlatformFromArtifactDirName('qa-report-macOS-22818315023')
+    ).toBe('macos')
+    expect(
+      extractPlatformFromArtifactDirName('qa-report-Linux-22818315023')
+    ).toBe('linux')
+  })
+
+  it('falls back to slugifying unknown directory names', () => {
+    expect(extractPlatformFromArtifactDirName('custom platform run')).toBe(
+      'custom-platform-run'
+    )
+  })
+})
+
+describe('pickLatestVideosByPlatform', () => {
+  it('keeps only the latest candidate per platform', () => {
+    const selected = pickLatestVideosByPlatform([
+      {
+        platformName: 'windows',
+        videoPath: '/tmp/windows-old.mp4',
+        mtimeMs: 100
+      },
+      {
+        platformName: 'windows',
+        videoPath: '/tmp/windows-new.mp4',
+        mtimeMs: 200
+      },
+      {
+        platformName: 'linux',
+        videoPath: '/tmp/linux.mp4',
+        mtimeMs: 150
+      }
+    ])
+
+    expect(selected).toEqual([
+      {
+        platformName: 'linux',
+        videoPath: '/tmp/linux.mp4',
+        mtimeMs: 150
+      },
+      {
+        platformName: 'windows',
+        videoPath: '/tmp/windows-new.mp4',
+        mtimeMs: 200
+      }
+    ])
+  })
+})
+
+describe('extractOutputText', () => {
+  it('reads output_text when available', () => {
+    expect(extractOutputText({ output_text: 'hello world' })).toBe(
+      'hello world'
+    )
+  })
+
+  it('reads nested output text chunks', () => {
+    const responseBody = {
+      output: [
+        {
+          content: [
+            { type: 'output_text', text: 'first block' },
+            { type: 'refusal', refusal: 'not relevant' }
+          ]
+        },
+        {
+          content: [{ type: 'output_text', text: 'second block' }]
+        }
+      ]
+    }
+
+    expect(extractOutputText(responseBody)).toBe('first block\n\nsecond block')
+  })
+
+  it('returns null when there is no usable text', () => {
+    expect(
+      extractOutputText({ output: [{ content: [{ type: 'tool_call' }] }] })
+    ).toBeNull()
+  })
+})
+
+describe('selectVideoCandidateByFile', () => {
+  it('selects a single candidate by artifacts-relative path', () => {
+    const selected = selectVideoCandidateByFile(
+      [
+        {
+          platformName: 'windows',
+          videoPath: '/tmp/qa-artifacts/qa-report-Windows-1/qa-session.mp4',
+          mtimeMs: 100
+        },
+        {
+          platformName: 'linux',
+          videoPath: '/tmp/qa-artifacts/qa-report-Linux-1/qa-session.mp4',
+          mtimeMs: 200
+        }
+      ],
+      {
+        artifactsDir: '/tmp/qa-artifacts',
+        videoFile: 'qa-report-Linux-1/qa-session.mp4'
+      }
+    )
+
+    expect(selected).toEqual({
+      platformName: 'linux',
+      videoPath: '/tmp/qa-artifacts/qa-report-Linux-1/qa-session.mp4',
+      mtimeMs: 200
+    })
+  })
+
+  it('throws when basename matches multiple videos', () => {
+    expect(() =>
+      selectVideoCandidateByFile(
+        [
+          {
+            platformName: 'windows',
+            videoPath: '/tmp/qa-artifacts/qa-report-Windows-1/qa-session.mp4',
+            mtimeMs: 100
+          },
+          {
+            platformName: 'linux',
+            videoPath: '/tmp/qa-artifacts/qa-report-Linux-1/qa-session.mp4',
+            mtimeMs: 200
+          }
+        ],
+        {
+          artifactsDir: '/tmp/qa-artifacts',
+          videoFile: 'qa-session.mp4'
+        }
+      )
+    ).toThrow('matched 2 videos')
+  })
+
+  it('throws when there is no matching video', () => {
+    expect(() =>
+      selectVideoCandidateByFile(
+        [
+          {
+            platformName: 'windows',
+            videoPath: '/tmp/qa-artifacts/qa-report-Windows-1/qa-session.mp4',
+            mtimeMs: 100
+          }
+        ],
+        {
+          artifactsDir: '/tmp/qa-artifacts',
+          videoFile: 'qa-report-macOS-1/qa-session.mp4'
+        }
+      )
+    ).toThrow('No video matched')
+  })
+
+  it('throws when video file is missing', () => {
+    expect(() =>
+      selectVideoCandidateByFile(
+        [
+          {
+            platformName: 'windows',
+            videoPath: '/tmp/qa-artifacts/qa-report-Windows-1/qa-session.mp4',
+            mtimeMs: 100
+          }
+        ],
+        {
+          artifactsDir: '/tmp/qa-artifacts',
+          videoFile: '   '
+        }
+      )
+    ).toThrow('--video-file is required')
+  })
+})

scripts/qa-video-review.ts

@@ -0,0 +1,795 @@
+#!/usr/bin/env tsx
+import { spawn } from 'node:child_process'
+import {
+  mkdtemp,
+  mkdir,
+  readdir,
+  readFile,
+  rm,
+  stat,
+  writeFile
+} from 'node:fs/promises'
+import { basename, dirname, join, relative, resolve } from 'node:path'
+import { fileURLToPath } from 'node:url'
+
+import { globSync } from 'glob'
+
+interface CliOptions {
+  artifactsDir: string
+  videoFile: string
+  outputDir: string
+  model: string
+  minIntervalSeconds: number
+  maxFrames: number
+  requestTimeoutMs: number
+  dryRun: boolean
+  keepFrames: boolean
+}
+
+interface VideoCandidate {
+  platformName: string
+  videoPath: string
+  mtimeMs: number
+}
+
+interface ExtractedFrame {
+  index: number
+  timestampSeconds: number
+  filePath: string
+  dataUrl: string
+}
+
+type ResponseInputContent =
+  | { type: 'input_text'; text: string }
+  | { type: 'input_image'; image_url: string }
+
+interface ResponseInputMessage {
+  role: 'system' | 'user'
+  content: ResponseInputContent[]
+}
+
+interface ResponsesCreatePayload {
+  model: string
+  input: ResponseInputMessage[]
+  max_output_tokens: number
+}
+
+interface OpenAIReviewRequest {
+  apiKey: string
+  baseUrl: string
+  model: string
+  platformName: string
+  videoPath: string
+  frames: ExtractedFrame[]
+  samplingIntervalSeconds: number
+  timeoutMs: number
+}
+
+const DEFAULT_OPTIONS: CliOptions = {
+  artifactsDir: './tmp/qa-artifacts',
+  videoFile: '',
+  outputDir: './tmp',
+  model: 'gpt-4o',
+  minIntervalSeconds: 5,
+  maxFrames: 36,
+  requestTimeoutMs: 300_000,
+  dryRun: false,
+  keepFrames: false
+}
+
+const USAGE = `Usage:
+  pnpm exec tsx scripts/qa-video-review.ts [options]
+
+Options:
+  --artifacts-dir <path>        Artifacts root directory
+                                 (default: ./tmp/qa-artifacts)
+  --video-file <name-or-path>   Video file to analyze (required)
+                                 (supports basename or relative/absolute path)
+  --output-dir <path>           Output directory for markdown reports
+                                 (default: ./tmp)
+  --model <name>                OpenAI model
+                                 (default: gpt-4o)
+  --min-interval-seconds <n>    Minimum frame sampling interval in seconds
+                                 (default: 5)
+  --max-frames <n>              Max frames analyzed per video
+                                 (default: 36)
+  --request-timeout-ms <n>      Request timeout in milliseconds
+                                 (default: 300000)
+  --dry-run                     Discover videos and output targets only
+  --keep-frames                 Keep extracted frames in ./tmp for inspection
+  --help                        Show this help text
+
+Environment:
+  OPENAI_API_KEY                Required unless --dry-run
+  OPENAI_BASE_URL               Optional override for API base URL
+`
+
+function parsePositiveNumber(rawValue: string, flagName: string): number {
+  const parsedValue = Number(rawValue)
+  if (!Number.isFinite(parsedValue) || parsedValue <= 0) {
+    throw new Error(`Invalid value for ${flagName}: "${rawValue}"`)
+  }
+  return parsedValue
+}
+
+function parsePositiveInteger(rawValue: string, flagName: string): number {
+  const parsedValue = Number.parseInt(rawValue, 10)
+  if (!Number.isInteger(parsedValue) || parsedValue <= 0) {
+    throw new Error(`Invalid value for ${flagName}: "${rawValue}"`)
+  }
+  return parsedValue
+}
+
+function parseCliOptions(args: string[]): CliOptions {
+  const options: CliOptions = { ...DEFAULT_OPTIONS }
+
+  for (let index = 0; index < args.length; index += 1) {
+    const argument = args[index]
+    const nextValue = args[index + 1]
+    const requireValue = (flagName: string): string => {
+      if (!nextValue || nextValue.startsWith('--')) {
+        throw new Error(`Missing value for ${flagName}`)
+      }
+      index += 1
+      return nextValue
+    }
+
+    if (argument === '--help') {
+      process.stdout.write(USAGE)
+      process.exit(0)
+    }
+
+    if (argument === '--artifacts-dir') {
+      options.artifactsDir = requireValue(argument)
+      continue
+    }
+
+    if (argument === '--video-file') {
+      options.videoFile = requireValue(argument)
+      continue
+    }
+
+    if (argument === '--output-dir') {
+      options.outputDir = requireValue(argument)
+      continue
+    }
+
+    if (argument === '--model') {
+      options.model = requireValue(argument)
+      continue
+    }
+
+    if (argument === '--min-interval-seconds') {
+      options.minIntervalSeconds = parsePositiveNumber(
+        requireValue(argument),
+        argument
+      )
+      continue
+    }
+
+    if (argument === '--max-frames') {
+      options.maxFrames = parsePositiveInteger(requireValue(argument), argument)
+      continue
+    }
+
+    if (argument === '--request-timeout-ms') {
+      options.requestTimeoutMs = parsePositiveInteger(
+        requireValue(argument),
+        argument
+      )
+      continue
+    }
+
+    if (argument === '--dry-run') {
+      options.dryRun = true
+      continue
+    }
+
+    if (argument === '--keep-frames') {
+      options.keepFrames = true
+      continue
+    }
+
+    throw new Error(`Unknown argument: ${argument}`)
+  }
+
+  return options
+}
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === 'object' && value !== null
+}
+
+function normalizePlatformName(value: string): string {
+  const slug = value
+    .trim()
+    .toLowerCase()
+    .replace(/[^a-z0-9]+/g, '-')
+    .replace(/^-+|-+$/g, '')
+
+  return slug.length > 0 ? slug : 'unknown-platform'
+}
+
+export function extractPlatformFromArtifactDirName(dirName: string): string {
+  const matchedValue = dirName.match(/^qa-report-(.+?)(?:-\d+)?$/i)?.[1]
+  return normalizePlatformName(matchedValue ?? dirName)
+}
+
+function extractPlatformFromVideoPath(videoPath: string): string {
+  const artifactDirName = basename(dirname(videoPath))
+  return extractPlatformFromArtifactDirName(artifactDirName)
+}
+
+export function pickLatestVideosByPlatform(
+  candidates: VideoCandidate[]
+): VideoCandidate[] {
+  const latestByPlatform = new Map<string, VideoCandidate>()
+
+  for (const candidate of candidates) {
+    const current = latestByPlatform.get(candidate.platformName)
+    if (!current || candidate.mtimeMs > current.mtimeMs) {
+      latestByPlatform.set(candidate.platformName, candidate)
+    }
+  }
+
+  return [...latestByPlatform.values()].sort((a, b) =>
+    a.platformName.localeCompare(b.platformName)
+  )
+}
+
+function toProjectRelativePath(targetPath: string): string {
+  const relativePath = relative(process.cwd(), targetPath)
+  if (relativePath.startsWith('.')) {
+    return relativePath
+  }
+  return `./${relativePath}`
+}
+
+function errorToString(error: unknown): string {
+  return error instanceof Error ? error.message : String(error)
+}
+
+function normalizePathForMatch(value: string): string {
+  return value.replaceAll('\\', '/').replace(/^\.\/+/, '')
+}
+
+export function selectVideoCandidateByFile(
+  candidates: VideoCandidate[],
+  options: { artifactsDir: string; videoFile: string }
+): VideoCandidate {
+  const requestedValue = options.videoFile.trim()
+  if (requestedValue.length === 0) {
+    throw new Error('--video-file is required')
+  }
+
+  const artifactsRoot = resolve(options.artifactsDir)
+  const requestedAbsolutePath = resolve(requestedValue)
+  const requestedPathKey = normalizePathForMatch(requestedValue)
+
+  const matches = candidates.filter((candidate) => {
+    const candidateAbsolutePath = resolve(candidate.videoPath)
+    if (candidateAbsolutePath === requestedAbsolutePath) {
+      return true
+    }
+
+    const candidateBaseName = basename(candidate.videoPath)
+    if (candidateBaseName === requestedValue) {
+      return true
+    }
+
+    const relativeToCwd = normalizePathForMatch(
+      relative(process.cwd(), candidateAbsolutePath)
+    )
+    if (relativeToCwd === requestedPathKey) {
+      return true
+    }
+
+    const relativeToArtifacts = normalizePathForMatch(
+      relative(artifactsRoot, candidateAbsolutePath)
+    )
+    return relativeToArtifacts === requestedPathKey
+  })
+
+  if (matches.length === 1) {
+    return matches[0]
+  }
+
+  if (matches.length === 0) {
+    const availableVideos = candidates.map((candidate) =>
+      toProjectRelativePath(candidate.videoPath)
+    )
+    throw new Error(
+      [
+        `No video matched --video-file "${options.videoFile}".`,
+        'Available videos:',
+        ...availableVideos.map((videoPath) => `- ${videoPath}`)
+      ].join('\n')
+    )
+  }
+
+  throw new Error(
+    [
+      `--video-file "${options.videoFile}" matched ${matches.length} videos.`,
+      'Please pass a more specific path.',
+      ...matches.map((match) => `- ${toProjectRelativePath(match.videoPath)}`)
+    ].join('\n')
+  )
+}
+
+async function runCommand(command: string, args: string[]): Promise<string> {
+  return await new Promise<string>((resolvePromise, rejectPromise) => {
+    const child = spawn(command, args, {
+      stdio: ['ignore', 'pipe', 'pipe']
+    })
+
+    let stdout = ''
+    let stderr = ''
+
+    child.stdout.on('data', (chunk: Buffer | string) => {
+      stdout += chunk.toString()
+    })
+
+    child.stderr.on('data', (chunk: Buffer | string) => {
+      stderr += chunk.toString()
+    })
+
+    child.on('error', (error) => {
+      rejectPromise(error)
+    })
+
+    child.on('close', (code) => {
+      if (code === 0) {
+        resolvePromise(stdout)
+        return
+      }
+
+      const details = stderr.trim() || stdout.trim() || 'unknown failure'
+      rejectPromise(
+        new Error(
+          `${command} ${args.join(' ')} failed with code ${String(code)}: ${details}`
+        )
+      )
+    })
+  })
+}
+
+async function ensureBinariesAvailable(): Promise<void> {
+  await runCommand('ffmpeg', ['-version'])
+  await runCommand('ffprobe', ['-version'])
+}
+
+async function getVideoDurationSeconds(
+  videoPath: string
+): Promise<number | null> {
+  try {
+    const stdout = await runCommand('ffprobe', [
+      '-v',
+      'error',
+      '-show_entries',
+      'format=duration',
+      '-of',
+      'default=noprint_wrappers=1:nokey=1',
+      videoPath
+    ])
+    const durationSeconds = Number.parseFloat(stdout.trim())
+    if (!Number.isFinite(durationSeconds) || durationSeconds <= 0) {
+      return null
+    }
+    return durationSeconds
+  } catch {
+    return null
+  }
+}
+
+function computeSamplingIntervalSeconds(
+  durationSeconds: number | null,
+  minIntervalSeconds: number,
+  maxFrames: number
+): number {
+  if (durationSeconds === null) {
+    return minIntervalSeconds
+  }
+
+  const fullCoverageInterval = durationSeconds / maxFrames
+  return Math.max(minIntervalSeconds, fullCoverageInterval)
+}
+
+async function collectVideoCandidates(
+  artifactsDir: string
+): Promise<VideoCandidate[]> {
+  const absoluteArtifactsDir = resolve(artifactsDir)
+  const videoPaths = globSync('**/qa-session.mp4', {
+    cwd: absoluteArtifactsDir,
+    absolute: true,
+    nodir: true
+  }).sort()
+
+  const candidates = await Promise.all(
+    videoPaths.map(async (videoPath) => {
+      const videoStat = await stat(videoPath)
+      return {
+        platformName: extractPlatformFromVideoPath(videoPath),
+        videoPath,
+        mtimeMs: videoStat.mtimeMs
+      }
+    })
+  )
+
+  return candidates
+}
+
+async function extractFramesFromVideo(
+  videoPath: string,
+  stagingDir: string,
+  minIntervalSeconds: number,
+  maxFrames: number
+): Promise<{ frames: ExtractedFrame[]; samplingIntervalSeconds: number }> {
+  const durationSeconds = await getVideoDurationSeconds(videoPath)
+  const samplingIntervalSeconds = computeSamplingIntervalSeconds(
+    durationSeconds,
+    minIntervalSeconds,
+    maxFrames
+  )
+  const outputPattern = join(stagingDir, 'frame-%03d.jpg')
+
+  await runCommand('ffmpeg', [
+    '-hide_banner',
+    '-loglevel',
+    'error',
+    '-y',
+    '-i',
+    videoPath,
+    '-vf',
+    `fps=1/${samplingIntervalSeconds},scale=1024:-2:force_original_aspect_ratio=decrease`,
+    '-frames:v',
+    String(maxFrames),
+    outputPattern
+  ])
+
+  const frameNames = (await readdir(stagingDir))
+    .filter((name) => name.startsWith('frame-') && name.endsWith('.jpg'))
+    .sort()
+
+  if (frameNames.length === 0) {
+    throw new Error(`No frames were extracted from ${videoPath}`)
+  }
+
+  const frames = await Promise.all(
+    frameNames.map(async (frameName, index) => {
+      const framePath = join(stagingDir, frameName)
+      const buffer = await readFile(framePath)
+      return {
+        index: index + 1,
+        timestampSeconds: Number((index * samplingIntervalSeconds).toFixed(2)),
+        filePath: framePath,
+        dataUrl: `data:image/jpeg;base64,${buffer.toString('base64')}`
+      }
+    })
+  )
+
+  return { frames, samplingIntervalSeconds }
+}
+
+function buildReviewMessages(
+  request: OpenAIReviewRequest
+): ResponseInputMessage[] {
+  const systemPrompt = [
+    'You are a senior QA engineer reviewing a UI test session.',
+    'Report only concrete, visible problems and avoid speculation.',
+    'If confidence is low, mark it explicitly.'
+  ].join(' ')
+
+  const introPrompt = [
+    `Review sampled frames from a QA session for platform "${request.platformName}".`,
+    `Source video: ${toProjectRelativePath(request.videoPath)}.`,
+    `Frames are presented in chronological order at ~${request.samplingIntervalSeconds.toFixed(2)}s intervals.`,
+    'Focus on UI regressions, broken states, visual glitches, unreadable text, missing labels/i18n, and clear workflow failures.',
+    'Return markdown with these sections exactly:',
+    '## Summary',
+    '## Confirmed Issues',
+    '## Possible Issues (Needs Human Verification)',
+    '## Overall Risk',
+    'Under Confirmed Issues include a markdown table with columns:',
+    'Severity | Timestamp (s) | Issue | Evidence | Confidence | Suggested Fix'
+  ].join('\n')
+
+  const userContent: ResponseInputContent[] = [
+    { type: 'input_text', text: introPrompt }
+  ]
+
+  for (const frame of request.frames) {
+    userContent.push({
+      type: 'input_text',
+      text: `Frame ${frame.index} at approximately ${frame.timestampSeconds}s`
+    })
+    userContent.push({
+      type: 'input_image',
+      image_url: frame.dataUrl
+    })
+  }
+
+  return [
+    {
+      role: 'system',
+      content: [{ type: 'input_text', text: systemPrompt }]
+    },
+    {
+      role: 'user',
+      content: userContent
+    }
+  ]
+}
+
+function normalizeBaseUrl(baseUrl: string): string {
+  return baseUrl.endsWith('/') ? baseUrl.slice(0, -1) : baseUrl
+}
+
+function extractApiErrorMessage(responseBody: unknown): string {
+  if (!isRecord(responseBody)) {
+    return 'Unknown API error'
+  }
+
+  const errorValue = responseBody.error
+  if (!isRecord(errorValue)) {
+    return 'Unknown API error'
+  }
+
+  const messageValue = errorValue.message
+  if (typeof messageValue !== 'string' || messageValue.length === 0) {
+    return 'Unknown API error'
+  }
+
+  return messageValue
+}
+
+export function extractOutputText(responseBody: unknown): string | null {
+  if (!isRecord(responseBody)) {
+    return null
+  }
+
+  const topLevelOutputText = responseBody.output_text
+  if (
+    typeof topLevelOutputText === 'string' &&
+    topLevelOutputText.trim().length > 0
+  ) {
+    return topLevelOutputText.trim()
+  }
+
+  const outputValue = responseBody.output
+  if (!Array.isArray(outputValue)) {
+    return null
+  }
+
+  const collectedParts: string[] = []
+  for (const outputItem of outputValue) {
+    if (!isRecord(outputItem)) {
+      continue
+    }
+
+    const contentValue = outputItem.content
+    if (!Array.isArray(contentValue)) {
+      continue
+    }
+
+    for (const contentItem of contentValue) {
+      if (!isRecord(contentItem)) {
+        continue
+      }
+
+      if (contentItem.type !== 'output_text') {
+        continue
+      }
+
+      const textValue = contentItem.text
+      if (typeof textValue === 'string' && textValue.trim().length > 0) {
+        collectedParts.push(textValue.trim())
+      }
+    }
+  }
+
+  if (collectedParts.length === 0) {
+    return null
+  }
+
+  return collectedParts.join('\n\n')
+}
+
+async function requestOpenAIReview(
+  request: OpenAIReviewRequest
+): Promise<string> {
+  const payload: ResponsesCreatePayload = {
+    model: request.model,
+    input: buildReviewMessages(request),
+    max_output_tokens: 3_000
+  }
+
+  const controller = new AbortController()
+  const timeoutHandle = setTimeout(() => controller.abort(), request.timeoutMs)
+
+  try {
+    const response = await fetch(
+      `${normalizeBaseUrl(request.baseUrl)}/responses`,
+      {
+        method: 'POST',
+        headers: {
+          Authorization: `Bearer ${request.apiKey}`,
+          'Content-Type': 'application/json'
+        },
+        body: JSON.stringify(payload),
+        signal: controller.signal
+      }
+    )
+
+    const rawResponse = await response.text()
+    let responseBody: unknown
+    try {
+      responseBody = JSON.parse(rawResponse)
+    } catch {
+      throw new Error(
+        `OpenAI API returned a non-JSON response: ${rawResponse.slice(0, 400)}`
+      )
+    }
+
+    if (!response.ok) {
+      throw new Error(
+        `OpenAI API request failed (${response.status}): ${extractApiErrorMessage(responseBody)}`
+      )
+    }
+
+    const reviewText = extractOutputText(responseBody)
+    if (!reviewText) {
+      throw new Error('OpenAI API returned no output text')
+    }
+
+    return reviewText
+  } finally {
+    clearTimeout(timeoutHandle)
+  }
+}
+
+function buildReportMarkdown(input: {
+  platformName: string
+  model: string
+  videoPath: string
+  frames: ExtractedFrame[]
+  samplingIntervalSeconds: number
+  reviewText: string
+}): string {
+  const header = [
+    `# ${input.platformName} QA Video Report`,
+    '',
+    `- Generated at: ${new Date().toISOString()}`,
+    `- Model: \`${input.model}\``,
+    `- Source video: \`${toProjectRelativePath(input.videoPath)}\``,
+    `- Frames analyzed: ${input.frames.length}`,
+    `- Sampling interval: ${input.samplingIntervalSeconds.toFixed(2)}s`,
+    '',
+    '## AI Review',
+    ''
+  ].join('\n')
+
+  return `${header}${input.reviewText.trim()}\n`
+}
+
+async function reviewVideo(
+  video: VideoCandidate,
+  options: CliOptions,
+  apiKey: string,
+  baseUrl: string
+): Promise<void> {
+  const stagingRoot = resolve('./tmp')
+  await mkdir(stagingRoot, { recursive: true })
+  const stagingDir = await mkdtemp(join(stagingRoot, 'qa-video-review-'))
+
+  try {
+    process.stdout.write(
+      `[${video.platformName}] Extracting frames from ${toProjectRelativePath(video.videoPath)}\n`
+    )
+
+    const { frames, samplingIntervalSeconds } = await extractFramesFromVideo(
+      video.videoPath,
+      stagingDir,
+      options.minIntervalSeconds,
+      options.maxFrames
+    )
+
+    process.stdout.write(
+      `[${video.platformName}] Sending ${frames.length} frame(s) to ${options.model}\n`
+    )
+
+    const reviewText = await requestOpenAIReview({
+      apiKey,
+      baseUrl,
+      model: options.model,
+      platformName: video.platformName,
+      videoPath: video.videoPath,
+      frames,
+      samplingIntervalSeconds,
+      timeoutMs: options.requestTimeoutMs
+    })
+
+    const outputPath = resolve(
+      options.outputDir,
+      `${video.platformName}-qa-video-report.md`
+    )
+    const reportMarkdown = buildReportMarkdown({
+      platformName: video.platformName,
+      model: options.model,
+      videoPath: video.videoPath,
+      frames,
+      samplingIntervalSeconds,
+      reviewText
+    })
+
+    await mkdir(dirname(outputPath), { recursive: true })
+    await writeFile(outputPath, reportMarkdown, 'utf-8')
+
+    process.stdout.write(
+      `[${video.platformName}] Wrote ${toProjectRelativePath(outputPath)}\n`
+    )
+
+    if (options.keepFrames) {
+      const stagedFrames = (await readdir(stagingDir)).filter((name) =>
+        name.endsWith('.jpg')
+      ).length
+      process.stdout.write(
+        `[${video.platformName}] Kept ${stagedFrames} frame(s) in ${toProjectRelativePath(stagingDir)}\n`
+      )
+      return
+    }
+  } finally {
+    if (!options.keepFrames) {
+      await rm(stagingDir, { recursive: true, force: true })
+    }
+  }
+}
+
+function isExecutedAsScript(metaUrl: string): boolean {
+  const modulePath = fileURLToPath(metaUrl)
+  const scriptPath = process.argv[1] ? resolve(process.argv[1]) : ''
+  return modulePath === scriptPath
+}
+
+async function main(): Promise<void> {
+  const options = parseCliOptions(process.argv.slice(2))
+  const candidates = await collectVideoCandidates(options.artifactsDir)
+
+  if (candidates.length === 0) {
+    process.stdout.write(
+      `No qa-session.mp4 files found under ${toProjectRelativePath(resolve(options.artifactsDir))}\n`
+    )
+    return
+  }
+
+  const selectedVideo = selectVideoCandidateByFile(candidates, {
+    artifactsDir: options.artifactsDir,
+    videoFile: options.videoFile
+  })
+
+  process.stdout.write(
+    `Selected ${selectedVideo.platformName}: ${toProjectRelativePath(selectedVideo.videoPath)}\n`
+  )
+
+  if (options.dryRun) {
+    process.stdout.write('\nDry run mode enabled, no API calls were made.\n')
+    return
+  }
+
+  const apiKey = process.env.OPENAI_API_KEY
+  if (!apiKey) {
+    throw new Error('OPENAI_API_KEY is required unless --dry-run is set')
+  }
+
+  await ensureBinariesAvailable()
+
+  const baseUrl = process.env.OPENAI_BASE_URL ?? 'https://api.openai.com/v1'
+  await reviewVideo(selectedVideo, options, apiKey, baseUrl)
+}
+
+if (isExecutedAsScript(import.meta.url)) {
+  void main().catch((error: unknown) => {
+    const message = errorToString(error)
+    process.stderr.write(`qa-video-review failed: ${message}\n`)
+    process.exit(1)
+  })
+}