[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M76YY2ZXYTWTNE3B08F63X08":3,"skill-same:01M76YY2ZXYTWTNE3B08F63X08":65},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":42,"commitSha":43,"skillDirectory":44,"forksCount":45,"repositoryHtmlUrl":46,"repositoryHomepageUrl":6,"ownerHtmlUrl":47,"ownerBlogUrl":48,"ownerDescription":49,"updatedAt":50,"install":51,"resources":64},"01M76YY2ZXYTWTNE3B08F63X08","youtube-video-analyzer",null,"Multimodal YouTube video analysis through both audio (transcript) and visual (frame extraction + image analysis) channels. Especially powerful for HowTo videos, tutorials, demos, and explainer videos where what is SHOWN (screenshots, UI demos, diagrams, code, physical actions) is just as important as what is SAID. Use this skill whenever a user wants to analyze, summarize, or create step-by-step guides from YouTube videos, or when they share a YouTube URL and want to understand what happens in the video. Triggers on requests like \"Analyze this YouTube video\", \"Create a step-by-step guide from this video\", \"What does this video show?\", \"Summarize this tutorial\", or any YouTube URL shared with analysis intent.","javimosch\u002Fopen-claw-skills","javimosch","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F12900508?v=4","01M6D3F2AS6PV75ET9PZ1QGP8J","Javier Leandro Arancibia","01M75B8NCTAJG9WH5YNZH7C3D8",0,"MIT","https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fsdrabent\u002Fyoutube-video-analyzer\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: youtube-video-analyzer\ndescription: >\n  Multimodal YouTube video analysis through both audio (transcript) and visual (frame extraction + image analysis) channels.\n  Especially powerful for HowTo videos, tutorials, demos, and explainer videos where what is SHOWN (screenshots, UI demos,\n  diagrams, code, physical actions) is just as important as what is SAID. Use this skill whenever a user wants to analyze,\n  summarize, or create step-by-step guides from YouTube videos, or when they share a YouTube URL and want to understand\n  what happens in the video. Triggers on requests like \"Analyze this YouTube video\", \"Create a step-by-step guide from\n  this video\", \"What does this video show?\", \"Summarize this tutorial\", or any YouTube URL shared with analysis intent.\nversion: 1.0.0\nmetadata:\n  openclaw:\n    requires:\n      bins:\n        - ffmpeg\n        - python3\n        - curl\n    emoji: \"🎬\"\n    os:\n      - linux\n      - macos\n    install:\n      - kind: uv\n        package: yt-dlp\n        bins: [yt-dlp]\n---\n\n# YouTube Video Analyzer — Multimodal\n\nThis skill performs deep analysis of YouTube videos through **both information channels**:\n- **Audio channel**: Transcript with timestamps (what is SAID)\n- **Visual channel**: Frame extraction + image analysis (what is SHOWN)\n\nMost YouTube skills only extract transcripts. This skill closes the gap by **synchronizing visual frames with spoken content**, enabling accurate step-by-step guides where \"click the blue button\" is matched with the actual screenshot showing which button.\n\n## Workflow Overview\n\n```\nYouTube URL\n    |\n    +---> 1. Get metadata (title, duration, video ID)\n    |\n    +---> 2. Extract transcript (yt-dlp --dump-json + curl)\n    |         -> Timestamped segments\n    |\n    +---> 3. Extract frames (yt-dlp + ffmpeg)\n    |         -> Keyframes at strategic intervals\n    |\n    +---> 4. Synchronize frames \u003C-> transcript\n    |         -> Match frames to spoken content by timestamp\n    |\n    +---> 5. Multimodal analysis\n              -> Read each frame image, combine with transcript\n              -> Generate structured output\n```\n\n## Step 1: Setup Working Directory\n\n```bash\nVIDEO_URL=\"\u003CYOUTUBE_URL>\"\nWORK_DIR=$(mktemp -d \u002Ftmp\u002Fyt-analysis-XXXXXX)\nmkdir -p \"$WORK_DIR\u002Fframes\"\n```\n\n## Step 2: Get Video Metadata\n\n```bash\nyt-dlp --print title --print duration --print id \"$VIDEO_URL\" 2>\u002Fdev\u002Fnull\n```\n\nThis returns three lines: title, duration in seconds, video ID. Store these for later use.\n\n## Step 3: Extract Transcript\n\n**IMPORTANT: Direct subtitle download via `--write-sub` frequently hits YouTube rate limits (HTTP 429).\nUse the reliable two-step method below instead.**\n\n### Step 3a: Get subtitle URL from video JSON\n\n```bash\nyt-dlp --dump-json \"$VIDEO_URL\" 2>\u002Fdev\u002Fnull | python3 -c \"\nimport json, sys\ndata = json.load(sys.stdin)\nauto = data.get('automatic_captions', {})\nsubs = data.get('subtitles', {})\n\n# Priority: manual subs > auto subs. Prefer user's language, fallback chain.\nfor source in [subs, auto]:\n    for lang in ['en', 'de', 'en-orig', 'fr', 'es']:\n        if lang in source:\n            for fmt in source[lang]:\n                if fmt.get('ext') == 'json3':\n                    print(fmt['url'])\n                    sys.exit(0)\n\n# Fallback: take first available auto-caption, get json3 URL\nfor lang in sorted(auto.keys()):\n    for fmt in auto[lang]:\n        if fmt.get('ext') == 'json3':\n            url = fmt['url']\n            # Remove translation param to get original language\n            import re\n            url = re.sub(r'&tlang=[^&]+', '', url)\n            print(url)\n            sys.exit(0)\n\nprint('NO_SUBS', file=sys.stderr)\nsys.exit(1)\n\" > \"$WORK_DIR\u002Fsub_url.txt\"\n```\n\n### Step 3b: Download and parse transcript\n\n```bash\ncurl -s \"$(cat \"$WORK_DIR\u002Fsub_url.txt\")\" -o \"$WORK_DIR\u002Ftranscript.json3\"\n```\n\nVerify it is valid JSON (not an HTML error page):\n\n```bash\nhead -c 20 \"$WORK_DIR\u002Ftranscript.json3\"\n# Should start with { — if it starts with \u003Chtml, retry after 10s sleep\n```\n\n### Step 3c: Parse json3 into readable timestamped segments\n\n```bash\npython3 -c \"\nimport json\n\nwith open('$WORK_DIR\u002Ftranscript.json3') as f:\n    data = json.load(f)\n\nfor event in data.get('events', []):\n    segs = event.get('segs', [])\n    if not segs:\n        continue\n    start_ms = event.get('tStartMs', 0)\n    duration_ms = event.get('dDurationMs', 0)\n    text = ''.join(s.get('utf8', '') for s in segs).strip()\n    if not text or text == '\\n':\n        continue\n    s = start_ms \u002F 1000\n    e = (start_ms + duration_ms) \u002F 1000\n    print(f'[{int(s\u002F\u002F60):02d}:{int(s%60):02d} - {int(e\u002F\u002F60):02d}:{int(e%60):02d}] {text}')\n\" > \"$WORK_DIR\u002Ftranscript.txt\"\n```\n\nRead `$WORK_DIR\u002Ftranscript.txt` to get the full transcript with timestamps.\n\n### Fallback: No transcript available\n\nIf no subtitles exist at all, inform the user and proceed with visual-only analysis.\n\n## Step 4: Download Video and Extract Frames\n\n### Step 4a: Download video (720p is sufficient for frame analysis)\n\n```bash\nyt-dlp -f \"bestvideo[height\u003C=720]+bestaudio\u002Fbest[height\u003C=720]\" \\\n       -o \"$WORK_DIR\u002Fvideo.mp4\" \"$VIDEO_URL\"\n```\n\n### Step 4b: Get exact duration\n\n```bash\nDURATION=$(ffprobe -v quiet -show_entries format=duration -of csv=p=0 \"$WORK_DIR\u002Fvideo.mp4\")\n```\n\n### Step 4c: Extract frames using adaptive interval strategy\n\nChoose interval based on video length:\n\n| Duration | Interval | Approx. Frames | Rationale |\n|----------|----------|-----------------|-----------|\n| \u003C 5 min | 10s | 20-30 | Dense enough for detailed analysis |\n| 5-20 min | 20s | 15-60 | Good balance of coverage vs. volume |\n| 20-60 min | 30-45s | 30-120 | Focus on key moments |\n| > 60 min | 60s | 60-120+ | Ask user if they want to focus on specific sections |\n\n```bash\n# Example for a 5-20 minute video (interval=20):\nffmpeg -i \"$WORK_DIR\u002Fvideo.mp4\" -vf \"fps=1\u002F20\" -q:v 3 \"$WORK_DIR\u002Fframes\u002Fframe_%04d.jpg\" 2>&1\n```\n\n**For scene-change-detection (software HowTos, UI demos):**\n\n```bash\nffmpeg -i \"$WORK_DIR\u002Fvideo.mp4\" \\\n       -vf \"select='gt(scene,0.3)',showinfo\" \\\n       -vsync vfr -q:v 3 \"$WORK_DIR\u002Fframes\u002Fscene_%04d.jpg\" 2>&1\n```\n\n### Step 4d: Calculate timestamps for each frame\n\nFor fixed-interval extraction: frame N has timestamp `(N-1) * interval` seconds.\n\n```\nframe_0001.jpg -> 0:00\nframe_0002.jpg -> 0:20\nframe_0003.jpg -> 0:40\n...\n```\n\n## Step 5: Synchronize Frames with Transcript\n\nFor each extracted frame:\n1. Calculate the frame's timestamp in seconds\n2. Find the transcript segment(s) covering that timestamp\n3. Create a synchronized pair: `{timestamp, transcript_text, frame_path}`\n\nThis is done mentally or via a simple lookup — no external script needed.\n\n## Step 6: Multimodal Analysis\n\n### Step 6a: Read and analyze each frame\n\nUse the `Read` tool (or `view` tool) to look at each frame image. For each frame, consider:\n\n- **UI elements**: Buttons, menus, dialogs, settings panels visible\n- **Text on screen**: Code, labels, error messages, URLs, terminal output\n- **Diagrams\u002Fgraphics**: Charts, flow diagrams, architecture drawings\n- **Physical actions**: Hand positions, tool usage (for physical HowTos)\n- **Changes**: What changed compared to the previous frame?\n\n### Step 6b: Synthesize both channels\n\nFor each key moment, combine audio and visual:\n\n```\nSegment [TIMESTAMP]:\n  SAID: \"Click the blue button in the top right\"\n  SHOWN: Settings page screenshot, blue \"Save\" button highlighted\n         in top-right corner, cursor pointing at it\n  SYNTHESIS: -> On the Settings page, click the blue \"Save\" button\n               in the top-right corner\n```\n\n### Step 6c: Identify visual-only information\n\nFlag moments where the visual channel provides information NOT present in audio:\n- Specific button names, menu paths, exact UI locations\n- Code that is shown but not read aloud\n- Error messages visible on screen\n- Before\u002Fafter comparisons\n\n## Output Formats\n\nGenerate the appropriate format based on the user's request:\n\n### Format A: Step-by-Step Guide (most common)\n\n```markdown\n# [Video Title] — Guide\n\n## Step 1: [Action] (00:15)\n[Description based on transcript + frame analysis]\n> Visual: [What the screen\u002Fimage shows at this point]\n\n## Step 2: [Action] (00:42)\n[...]\n```\n\n### Format B: Comprehensive Summary with Visual Anchors\n\n```markdown\n# [Video Title] — Summary\n\n## Overview\n[2-3 sentence summary of the entire video]\n\n## Key Sections\n\n### [Section Name] (00:00 - 02:30)\n[Summary of this section]\n- Key visual: [Description of what's shown]\n- Key quote: \"[Important spoken content]\"\n\n### [Section Name] (02:30 - 05:00)\n[...]\n\n## Key Takeaways\n- [Takeaway 1]\n- [Takeaway 2]\n```\n\n### Format C: Technical Detail Analysis\n\nSeparate analysis of both channels plus discrepancy detection:\n\n```markdown\n# [Video Title] — Technical Analysis\n\n## Audio Channel Analysis\n[What was said, key points, structure]\n\n## Visual Channel Analysis\n[What was shown, UI flows, code, diagrams]\n\n## Channel Synchronization\n[Where audio and visual complement each other]\n\n## Visual-Only Information\n[Important details only visible in frames, not mentioned in speech]\n```\n\n## Error Handling & Edge Cases\n\n| Problem | Solution |\n|---------|----------|\n| HTTP 429 on subtitle download | Use `--dump-json` method (Step 3a). If curl also gets blocked, wait 10-15 seconds and retry with different User-Agent |\n| No subtitles available at all | Proceed with visual-only analysis, inform user |\n| Original audio language not in auto-captions list | The original language is the source — auto-captions are translations. Remove `&tlang=XX` from any auto-caption URL to get the original |\n| `transcript.json3` contains HTML instead of JSON | YouTube returned an error page. Wait 10s, retry with: `curl -s --user-agent \"Mozilla\u002F5.0 (Windows NT 10.0; Win64; x64)\" \"$URL\"` |\n| Video > 60 min | Ask user if they want to focus on specific time ranges or chapters |\n| Poor video quality \u002F blurry frames | Extract more frames at tighter intervals to compensate |\n| Video is age-restricted or private | Inform user that the video cannot be accessed. Suggest using `--cookies-from-browser` if they have access |\n| yt-dlp download fails | Try alternative format: `-f \"best[height\u003C=720]\"` without separate audio+video streams |\n\n## Cleanup\n\nAfter analysis is complete, remove temporary files:\n\n```bash\nrm -rf \"$WORK_DIR\"\n```\n\n## Tips for Best Results\n\n- **Software HowTos**: Use scene-change detection — UI transitions create clear visual breaks\n- **Physical HowTos**: Use tighter frame intervals (10-15s) — movements are subtler\n- **Read the transcript first**: Identify \"interesting timestamps\" before extracting frames. Look for phrases like \"as you can see here\", \"let me show you\", \"on the screen\" — these signal important visual moments\n- **Context-aware frame analysis**: When analyzing a frame, always provide the transcript context. The speaker often explains what's about to be shown\n- **Batch frame reading**: Read frames in batches of 8-10 to maintain context across sequential frames and detect visual changes\n- **Always extract both channels in parallel**: Start the video download while processing the transcript to save time\n","# YouTube Video Analyzer — Multimodal\n\nThis skill performs deep analysis of YouTube videos through **both information channels**:\n- **Audio channel**: Transcript with timestamps (what is SAID)\n- **Visual channel**: Frame extraction + image analysis (what is SHOWN)\n\nMost YouTube skills only extract transcripts. This skill closes the gap by **synchronizing visual frames with spoken content**, enabling accurate step-by-step guides where \"click the blue button\" is matched with the actual screenshot showing which button.\n\n## Workflow Overview\n\n```\nYouTube URL\n    |\n    +---> 1. Get metadata (title, duration, video ID)\n    |\n    +---> 2. Extract transcript (yt-dlp --dump-json + curl)\n    |         -> Timestamped segments\n    |\n    +---> 3. Extract frames (yt-dlp + ffmpeg)\n    |         -> Keyframes at strategic intervals\n    |\n    +---> 4. Synchronize frames \u003C-> transcript\n    |         -> Match frames to spoken content by timestamp\n    |\n    +---> 5. Multimodal analysis\n              -> Read each frame image, combine with transcript\n              -> Generate structured output\n```\n\n## Step 1: Setup Working Directory\n\n```bash\nVIDEO_URL=\"\u003CYOUTUBE_URL>\"\nWORK_DIR=$(mktemp -d \u002Ftmp\u002Fyt-analysis-XXXXXX)\nmkdir -p \"$WORK_DIR\u002Fframes\"\n```\n\n## Step 2: Get Video Metadata\n\n```bash\nyt-dlp --print title --print duration --print id \"$VIDEO_URL\" 2>\u002Fdev\u002Fnull\n```\n\nThis returns three lines: title, duration in seconds, video ID. Store these for later use.\n\n## Step 3: Extract Transcript\n\n**IMPORTANT: Direct subtitle download via `--write-sub` frequently hits YouTube rate limits (HTTP 429).\nUse the reliable two-step method below instead.**\n\n### Step 3a: Get subtitle URL from video JSON\n\n```bash\nyt-dlp --dump-json \"$VIDEO_URL\" 2>\u002Fdev\u002Fnull | python3 -c \"\nimport json, sys\ndata = json.load(sys.stdin)\nauto = data.get('automatic_captions', {})\nsubs = data.get('subtitles', {})\n\n# Priority: manual subs > auto subs. Prefer user's language, fallback chain.\nfor source in [subs, auto]:\n    for lang in ['en', 'de', 'en-orig', 'fr', 'es']:\n        if lang in source:\n            for fmt in source[lang]:\n                if fmt.get('ext') == 'json3':\n                    print(fmt['url'])\n                    sys.exit(0)\n\n# Fallback: take first available auto-caption, get json3 URL\nfor lang in sorted(auto.keys()):\n    for fmt in auto[lang]:\n        if fmt.get('ext') == 'json3':\n            url = fmt['url']\n            # Remove translation param to get original language\n            import re\n            url = re.sub(r'&tlang=[^&]+', '', url)\n            print(url)\n            sys.exit(0)\n\nprint('NO_SUBS', file=sys.stderr)\nsys.exit(1)\n\" > \"$WORK_DIR\u002Fsub_url.txt\"\n```\n\n### Step 3b: Download and parse transcript\n\n```bash\ncurl -s \"$(cat \"$WORK_DIR\u002Fsub_url.txt\")\" -o \"$WORK_DIR\u002Ftranscript.json3\"\n```\n\nVerify it is valid JSON (not an HTML error page):\n\n```bash\nhead -c 20 \"$WORK_DIR\u002Ftranscript.json3\"\n# Should start with { — if it starts with \u003Chtml, retry after 10s sleep\n```\n\n### Step 3c: Parse json3 into readable timestamped segments\n\n```bash\npython3 -c \"\nimport json\n\nwith open('$WORK_DIR\u002Ftranscript.json3') as f:\n    data = json.load(f)\n\nfor event in data.get('events', []):\n    segs = event.get('segs', [])\n    if not segs:\n        continue\n    start_ms = event.get('tStartMs', 0)\n    duration_ms = event.get('dDurationMs', 0)\n    text = ''.join(s.get('utf8', '') for s in segs).strip()\n    if not text or text == '\\n':\n        continue\n    s = start_ms \u002F 1000\n    e = (start_ms + duration_ms) \u002F 1000\n    print(f'[{int(s\u002F\u002F60):02d}:{int(s%60):02d} - {int(e\u002F\u002F60):02d}:{int(e%60):02d}] {text}')\n\" > \"$WORK_DIR\u002Ftranscript.txt\"\n```\n\nRead `$WORK_DIR\u002Ftranscript.txt` to get the full transcript with timestamps.\n\n### Fallback: No transcript available\n\nIf no subtitles exist at all, inform the user and proceed with visual-only analysis.\n\n## Step 4: Download Video and Extract Frames\n\n### Step 4a: Download video (720p is sufficient for frame analysis)\n\n```bash\nyt-dlp -f \"bestvideo[height\u003C=720]+bestaudio\u002Fbest[height\u003C=720]\" \\\n       -o \"$WORK_DIR\u002Fvideo.mp4\" \"$VIDEO_URL\"\n```\n\n### Step 4b: Get exact duration\n\n```bash\nDURATION=$(ffprobe -v quiet -show_entries format=duration -of csv=p=0 \"$WORK_DIR\u002Fvideo.mp4\")\n```\n\n### Step 4c: Extract frames using adaptive interval strategy\n\nChoose interval based on video length:\n\n| Duration | Interval | Approx. Frames | Rationale |\n|----------|----------|-----------------|-----------|\n| \u003C 5 min | 10s | 20-30 | Dense enough for detailed analysis |\n| 5-20 min | 20s | 15-60 | Good balance of coverage vs. volume |\n| 20-60 min | 30-45s | 30-120 | Focus on key moments |\n| > 60 min | 60s | 60-120+ | Ask user if they want to focus on specific sections |\n\n```bash\n# Example for a 5-20 minute video (interval=20):\nffmpeg -i \"$WORK_DIR\u002Fvideo.mp4\" -vf \"fps=1\u002F20\" -q:v 3 \"$WORK_DIR\u002Fframes\u002Fframe_%04d.jpg\" 2>&1\n```\n\n**For scene-change-detection (software HowTos, UI demos):**\n\n```bash\nffmpeg -i \"$WORK_DIR\u002Fvideo.mp4\" \\\n       -vf \"select='gt(scene,0.3)',showinfo\" \\\n       -vsync vfr -q:v 3 \"$WORK_DIR\u002Fframes\u002Fscene_%04d.jpg\" 2>&1\n```\n\n### Step 4d: Calculate timestamps for each frame\n\nFor fixed-interval extraction: frame N has timestamp `(N-1) * interval` seconds.\n\n```\nframe_0001.jpg -> 0:00\nframe_0002.jpg -> 0:20\nframe_0003.jpg -> 0:40\n...\n```\n\n## Step 5: Synchronize Frames with Transcript\n\nFor each extracted frame:\n1. Calculate the frame's timestamp in seconds\n2. Find the transcript segment(s) covering that timestamp\n3. Create a synchronized pair: `{timestamp, transcript_text, frame_path}`\n\nThis is done mentally or via a simple lookup — no external script needed.\n\n## Step 6: Multimodal Analysis\n\n### Step 6a: Read and analyze each frame\n\nUse the `Read` tool (or `view` tool) to look at each frame image. For each frame, consider:\n\n- **UI elements**: Buttons, menus, dialogs, settings panels visible\n- **Text on screen**: Code, labels, error messages, URLs, terminal output\n- **Diagrams\u002Fgraphics**: Charts, flow diagrams, architecture drawings\n- **Physical actions**: Hand positions, tool usage (for physical HowTos)\n- **Changes**: What changed compared to the previous frame?\n\n### Step 6b: Synthesize both channels\n\nFor each key moment, combine audio and visual:\n\n```\nSegment [TIMESTAMP]:\n  SAID: \"Click the blue button in the top right\"\n  SHOWN: Settings page screenshot, blue \"Save\" button highlighted\n         in top-right corner, cursor pointing at it\n  SYNTHESIS: -> On the Settings page, click the blue \"Save\" button\n               in the top-right corner\n```\n\n### Step 6c: Identify visual-only information\n\nFlag moments where the visual channel provides information NOT present in audio:\n- Specific button names, menu paths, exact UI locations\n- Code that is shown but not read aloud\n- Error messages visible on screen\n- Before\u002Fafter comparisons\n\n## Output Formats\n\nGenerate the appropriate format based on the user's request:\n\n### Format A: Step-by-Step Guide (most common)\n\n```markdown\n# [Video Title] — Guide\n\n## Step 1: [Action] (00:15)\n[Description based on transcript + frame analysis]\n> Visual: [What the screen\u002Fimage shows at this point]\n\n## Step 2: [Action] (00:42)\n[...]\n```\n\n### Format B: Comprehensive Summary with Visual Anchors\n\n```markdown\n# [Video Title] — Summary\n\n## Overview\n[2-3 sentence summary of the entire video]\n\n## Key Sections\n\n### [Section Name] (00:00 - 02:30)\n[Summary of this section]\n- Key visual: [Description of what's shown]\n- Key quote: \"[Important spoken content]\"\n\n### [Section Name] (02:30 - 05:00)\n[...]\n\n## Key Takeaways\n- [Takeaway 1]\n- [Takeaway 2]\n```\n\n### Format C: Technical Detail Analysis\n\nSeparate analysis of both channels plus discrepancy detection:\n\n```markdown\n# [Video Title] — Technical Analysis\n\n## Audio Channel Analysis\n[What was said, key points, structure]\n\n## Visual Channel Analysis\n[What was shown, UI flows, code, diagrams]\n\n## Channel Synchronization\n[Where audio and visual complement each other]\n\n## Visual-Only Information\n[Important details only visible in frames, not mentioned in speech]\n```\n\n## Error Handling & Edge Cases\n\n| Problem | Solution |\n|---------|----------|\n| HTTP 429 on subtitle download | Use `--dump-json` method (Step 3a). If curl also gets blocked, wait 10-15 seconds and retry with different User-Agent |\n| No subtitles available at all | Proceed with visual-only analysis, inform user |\n| Original audio language not in auto-captions list | The original language is the source — auto-captions are translations. Remove `&tlang=XX` from any auto-caption URL to get the original |\n| `transcript.json3` contains HTML instead of JSON | YouTube returned an error page. Wait 10s, retry with: `curl -s --user-agent \"Mozilla\u002F5.0 (Windows NT 10.0; Win64; x64)\" \"$URL\"` |\n| Video > 60 min | Ask user if they want to focus on specific time ranges or chapters |\n| Poor video quality \u002F blurry frames | Extract more frames at tighter intervals to compensate |\n| Video is age-restricted or private | Inform user that the video cannot be accessed. Suggest using `--cookies-from-browser` if they have access |\n| yt-dlp download fails | Try alternative format: `-f \"best[height\u003C=720]\"` without separate audio+video streams |\n\n## Cleanup\n\nAfter analysis is complete, remove temporary files:\n\n```bash\nrm -rf \"$WORK_DIR\"\n```\n\n## Tips for Best Results\n\n- **Software HowTos**: Use scene-change detection — UI transitions create clear visual breaks\n- **Physical HowTos**: Use tighter frame intervals (10-15s) — movements are subtler\n- **Read the transcript first**: Identify \"interesting timestamps\" before extracting frames. Look for phrases like \"as you can see here\", \"let me show you\", \"on the screen\" — these signal important visual moments\n- **Context-aware frame analysis**: When analyzing a frame, always provide the transcript context. The speaker often explains what's about to be shown\n- **Batch frame reading**: Read frames in batches of 8-10 to maintain context across sequential frames and detect visual changes\n- **Always extract both channels in parallel**: Start the video download while processing the transcript to save time\n",{"name":5,"version":24,"metadata":25,"description":41},"1.0.0",{"openclaw":26},{"os":27,"emoji":30,"install":31,"requires":36},[28,29],"linux","macos","🎬",[32],{"bins":33,"kind":35,"package":34},[34],"yt-dlp","uv",{"bins":37},[38,39,40],"ffmpeg","python3","curl","Multimodal YouTube video analysis through both audio (transcript) and visual (frame extraction + image analysis) channels. Especially powerful for HowTo videos, tutorials, demos, and explainer videos where what is SHOWN (screenshots, UI demos, diagrams, code, physical actions) is just as important as what is SAID. Use this skill whenever a user wants to analyze, summarize, or create step-by-step guides from YouTube videos, or when they share a YouTube URL and want to understand what happens in the video. Triggers on requests like \"Analyze this YouTube video\", \"Create a step-by-step guide from this video\", \"What does this video show?\", \"Summarize this tutorial\", or any YouTube URL shared with analysis intent.\n","main","f9d3fe7cf9ac8f1ea63a0d96e98e1f510a3814db","skills\u002Fsdrabent\u002Fyoutube-video-analyzer",1,"https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills","https:\u002F\u002Fgithub.com\u002Fjavimosch","intrane.fr","Building robust, open-source infrastructure for the commons. AI-leveraged, self-hosted, built to last.","2026-06-07 03:34:58",{"source":8,"skillName":5,"commands":52,"prompt":62,"zipAvailable":63},[53,56,59],{"tool":54,"command":55},"npx","npx skills add javimosch\u002Fopen-claw-skills --skill youtube-video-analyzer",{"tool":57,"command":58},"bunx","bunx skills add javimosch\u002Fopen-claw-skills --skill youtube-video-analyzer",{"tool":60,"command":61},"pnpm","pnpm dlx skills add javimosch\u002Fopen-claw-skills --skill youtube-video-analyzer","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fsdrabent\u002Fyoutube-video-analyzer\u002FSKILL.md",true,[],[66,74,81,88,95,102],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":73},"01M778GGXG2VQ0YF31W2MKGMQZ","crowd-prompting","A marketplace where AI agents improve prompts, system instructions, tool descriptions, and other text-based content with domain expertise from real-world operations — and earn tokens for valuable contributions.","https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fzzadrian\u002Fcrowd-prompting\u002FSKILL.md",[],[],"2026-11-15 14:19:17",{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":73},"01M778G7BH8SBJ9FJ6C3G7CB1K","ceaser","Interact with the Ceaser privacy protocol on Base L2 using the ceaser-mcp MCP tools. This skill uses the ceaser-mcp npm package for ALL operations -- shield, unshield, note management, and protocol queries. All ceaser tool calls use CLI subcommands (npx -y ceaser-mcp \u003Csubcommand>). Ceaser is a privacy-preserving ETH wrapper using Noir\u002FUltraHonk zero-knowledge proofs on Base L2 (chain ID 8453).","https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fzyra-v21\u002Fceaser\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":73},"01M778FYCJQZWFKKX5FYMK3S47","ceaser-send","Fully automated private ETH transfer via Ceaser Protocol on Base L2 using the ceaser-mcp MCP tools. This skill uses the ceaser-mcp npm package for ALL operations -- all ceaser tool calls use CLI subcommands (npx -y ceaser-mcp \u003Csubcommand>). WARNING -- creates observable on-chain links between user wallet and hot wallet (see Privacy Warning). Generates an ephemeral hot wallet with BIP-39 mnemonic recovery, user funds it, agent signs and broadcasts the Shield TX automatically. Exactly one manual step required (funding the hot wallet with ETH). Uses Noir\u002FUltraHonk zero-knowledge proofs.","https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fzyra-v21\u002Fceaser-send\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":73},"01M778FQQNSEYA3HJKZBPQ4M9Z","dilemm-ai","Compete in DilemmAI, the prisoner's dilemma AI arena at dilemm.ai. Use when an OpenClaw agent wants to create an account, design and submit strategy prompts for their bot, enter matchmaking, analyze historical matches and leaderboards, and iterate on bot strategy. Requires headless browser automation (no API available yet).","https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fzyori\u002Fdilemm-ai\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":73},"01M778FGVZX7QA2TF9BFMM37FY","moldium","Post and manage content on the Moldium blog platform. Triggered by \"post to Moldium\", \"write a blog post\", \"publish an article\", etc.","https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fzyom45\u002Fmoldium\u002FSKILL.md",[],[],{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":73},"01M778F88D8PQ9MQ0KNW2TJ8ZH","arxiv-paper-reviews","Interact with arXiv Crawler API to fetch papers, read reviews, submit comments, search papers, and import papers. Use when working with arXiv papers, fetching paper lists by date\u002Fcategory\u002Finterest, viewing paper details with comments, submitting paper reviews, searching papers by title, or importing papers from arXiv URLs via API at http:\u002F\u002Fweakaccept.top:8000\u002F.","https:\u002F\u002Fgithub.com\u002Fjavimosch\u002Fopen-claw-skills\u002Fblob\u002Fmain\u002Fskills\u002Fzxrys\u002Farxiv-paper-reviews\u002FSKILL.md",[],[]]