[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74EMQCK2NBSPHJ6ESXZXN76":3,"skill-same:01M74EMQCK2NBSPHJ6ESXZXN76":54},{"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":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M74EMQCK2NBSPHJ6ESXZXN76","social-video-mining",null,"Extracts insights from short video (TikTok, Shorts, Reels, Douyin) without paid speech-to-text: yt-dlp, ffmpeg and vision over the frames. Use it when on-screen text carries the signal or when there is no transcript.","CarlosCaPe\u002Foctorato","CarlosCaPe","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F39838249?v=4","01M72WXX4ZRR8MBH1EP1QND2FQ","dataqbs","01M72WXX5NA7YCKRSJ54GW1GQ8",18,"MIT","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fsocial-video-mining\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: social-video-mining\ndescription: \"Extracts insights from short video (TikTok, Shorts, Reels, Douyin) without paid speech-to-text: yt-dlp, ffmpeg and vision over the frames. Use it when on-screen text carries the signal or when there is no transcript.\"\n---\n\n# Social Video Mining — Frames Over Audio\n\nTechnique skill, not a tool wrapper. Documents the workflow we used to mine OSS-replacement gems from a daily-tech-tip TikTok creator without burning OpenAI Whisper credits. The key insight: **tech short-form video puts the signal on screen, not in the audio** — project names, URLs, GitHub stats, license badges all appear as overlay text and screen recordings.\n\n## When to use\n\n- Curating skill candidates from a tech-creator channel (TikTok, YouTube Shorts, Reels, Douyin, Bilibili)\n- Verifying product claims from video reviews without watching dozens of videos\n- Building reference material from conference talks where transcripts are unavailable\n- Any \"video → structured knowledge\" task where audio transcription is blocked, expensive, or just not the highest-signal channel\n\n## When NOT to use\n\n- Audio IS the signal (interviews, podcasts, lectures) — use `transcribe` skill instead\n- Long-form video where the project name only appears once — frame sampling will miss it; consider audio transcription\n- Live streams or videos behind paywalls\n\n## Workflow\n\n```\n1. List      → yt-dlp --flat-playlist --dump-json --playlist-end N \u003Cchannel-url>\n                Extract: title, description, view_count, upload_date per video\n2. Triage    → filter titles\u002Fdescriptions for the signal you want\n                Sort by view_count (popularity = social validation)\n3. Download  → yt-dlp -q -o \"\u002Ftmp\u002F\u003Cprefix>_%(id)s.%(ext)s\" \u003Cvideo-url>\n4. Frames    → ffmpeg -ss \u003Csec> -i \u003Cvideo> -frames:v 1 -q:v 4 \u003Cframe>.jpg\n                Extract 2-3 frames per video at strategic timestamps:\n                  - t=2-5s : intro \u002F project name often appears\n                  - t=15-25s : middle \u002F detailed view of the tool\n                  - t=35-40s : outro \u002F CTA + URL often appears\n5. Vision    → Read each JPG — native multimodal sees:\n                  - Project names (top-left logos, browser tabs)\n                  - GitHub repo URLs (README banners, browser URL bar)\n                  - License badges (MIT \u002F Apache \u002F AGPL)\n                  - Star counts, version numbers\n                  - On-screen overlay text (Spanish captions, hashtags)\n6. Verify    → WebFetch or `gh search repos \"\u003Cname>\"` to confirm canonical URL exists\n                and license \u002F stars match what was on screen\n7. Cleanup   → rm -rf \u002Ftmp\u002F\u003Cprefix>_*  to release ~5MB per video\n```\n\n## Strategic timestamp picks\n\nMost tech-tip short-form videos follow this structure:\n\n| Timestamp | Typical content |\n|---|---|\n| 0-3s | Hook + caption with project pitch (\"AWS en local y gratis\") |\n| 3-10s | Project logo + landing page hero (\"Local AWS. Zero cost.\") |\n| 10-25s | Demo \u002F detail screen recording with on-screen features |\n| 25-35s | Stats \u002F license \u002F GitHub repo screenshot |\n| 35-45s | Outro CTA with URL + hashtags |\n\nSample frames at t=5, t=20, t=40 and you'll capture name, demo, and source URL in 90% of cases. Add t=12, t=32 for hard cases.\n\n## Cost profile (vs alternatives)\n\n| Approach | Tokens \u002F 1 video | API cost \u002F 1 video | Speed |\n|---|---|---|---|\n| WebFetch on TikTok URL | ~500 (bot-walled) | $0 | Fast, useless |\n| Browser scrape via agent-browser | ~5,000-10,000 | $0 | Slow, login-walled |\n| yt-dlp + OpenAI Whisper transcription | ~3,000 | ~$0.01 | Medium, needs API key |\n| **This skill (yt-dlp + frames + native vision)** | ~2,000 (3 frames × ~600 each) | **$0** | Fast, no API key |\n\nNative vision dominates for project-identification tasks. Transcription only wins when audio is the signal (which short-form tech video rarely is).\n\n## Permissions \u002F install prerequisites\n\n- `~\u002F.local\u002Fbin\u002Fyt-dlp` — install via standalone GitHub release (no PEP 668 conflict). Settings allowlist needs:\n  - `Bash(curl -fsSL https:\u002F\u002Fgithub.com\u002Fyt-dlp\u002Fyt-dlp\u002Freleases\u002F*)`\n  - `Bash(chmod +x \u002Fhome\u002F\u003Cuser>\u002F.local\u002Fbin\u002Fyt-dlp)`\n  - `Bash(~\u002F.local\u002Fbin\u002Fyt-dlp:*)`\n- `ffmpeg` — already present on most systems at `~\u002F.local\u002Fbin\u002Fffmpeg` or via `which ffmpeg`\n- `jq` — for parsing yt-dlp's JSON dumps\n- No external API keys required\n\n## Cleanup discipline\n\nAlways `rm -rf \u002Ftmp\u002F\u003Cprefix>_*` at the end of a mining run. Each video is ~5MB; a 10-video session = 50MB. Don't leave artifacts.\n\n## Limits \u002F when this fails\n\n- **Channel doesn't permit unauthenticated metadata access** — yt-dlp may need cookies; falls back to agent-browser\n- **Creator uses CapCut \u002F Premiere with minimal on-screen text** — name only spoken, never shown. Then audio transcription wins\n- **Branded text obscured by stylized fonts \u002F shaders** — vision misreads; use 2 frames to cross-check\n- **Project name is generic (\"API client\")** — vision identifies the screenshot but not the brand; WebFetch search step is essential\n\n## TikTok gotcha: do NOT add curl_cffi, and don't blindly upgrade yt-dlp\n\nVerified 2026-06-11 against a live TikTok pull. Two traps, both counter-intuitive:\n\n1. **curl_cffi impersonation BREAKS the TikTok extractor.** yt-dlp warns \"no impersonation target available\" on TikTok, which looks like the thing to fix. It is not. Inject `curl_cffi` and TikTok serves a challenge page with no rehydration JSON, so extraction dies with `ERROR: Unable to extract universal data for rehydration`. Plain urllib (no curl_cffi) gets the normal page and works. Leave that warning alone for TikTok.\n2. **yt-dlp versions after 2026.03.17 regressed TikTok** with the same rehydration error (2026.6.9 and master both fail, even without curl_cffi). Pin to a known-working build (`pipx runpip yt-dlp install \"yt-dlp==2026.3.17\"`) and only upgrade once an upstream TikTok fix lands. Test the actual target video after any yt-dlp change, never assume newer is safer.\n\nRecognition signal: `Unable to extract universal data for rehydration` on a TikTok URL = one of these two, not a network blip.\n\n## Anti-fabrication discipline (CRITICAL per CLAUDE.md)\n\nVision can hallucinate text it didn't actually see. **Every project name extracted by vision MUST be verified** via WebFetch or `gh search` before writing it into a brain skill or a client recommendation. The verification step is not optional. Workflow step 6 exists specifically to enforce this.\n\nIf verification fails (no repo found, name doesn't match search results), do NOT invent a URL or guess. Report \"vision saw X but no canonical repo found — manual verification needed\" and stop.\n\n## Related brain assets\n\n- `transcribe` — when audio IS the signal (interviews, lectures, podcasts)\n- `image-analyzer`, `gemini-vision` — for richer per-image analysis when frames need deeper interpretation\n- `agent-browser` — fallback when yt-dlp is blocked\n- `knowledge-corpus` — to organize what's mined into a queryable brain layer\n- Trend Researcher agent — natural consumer of mined social-video gems\n- Sister technique: the broader \"mine creator channels for OSS gems\" pattern that produced this skill, `claude-mem-persistent-memory`, `bruno-postman-alternative`, `coolify-self-hosted-paas`, `spacetimedb-backend-as-database`, and `vibevoice-elevenlabs-alternative` in a single research session\n","# Social Video Mining — Frames Over Audio\n\nTechnique skill, not a tool wrapper. Documents the workflow we used to mine OSS-replacement gems from a daily-tech-tip TikTok creator without burning OpenAI Whisper credits. The key insight: **tech short-form video puts the signal on screen, not in the audio** — project names, URLs, GitHub stats, license badges all appear as overlay text and screen recordings.\n\n## When to use\n\n- Curating skill candidates from a tech-creator channel (TikTok, YouTube Shorts, Reels, Douyin, Bilibili)\n- Verifying product claims from video reviews without watching dozens of videos\n- Building reference material from conference talks where transcripts are unavailable\n- Any \"video → structured knowledge\" task where audio transcription is blocked, expensive, or just not the highest-signal channel\n\n## When NOT to use\n\n- Audio IS the signal (interviews, podcasts, lectures) — use `transcribe` skill instead\n- Long-form video where the project name only appears once — frame sampling will miss it; consider audio transcription\n- Live streams or videos behind paywalls\n\n## Workflow\n\n```\n1. List      → yt-dlp --flat-playlist --dump-json --playlist-end N \u003Cchannel-url>\n                Extract: title, description, view_count, upload_date per video\n2. Triage    → filter titles\u002Fdescriptions for the signal you want\n                Sort by view_count (popularity = social validation)\n3. Download  → yt-dlp -q -o \"\u002Ftmp\u002F\u003Cprefix>_%(id)s.%(ext)s\" \u003Cvideo-url>\n4. Frames    → ffmpeg -ss \u003Csec> -i \u003Cvideo> -frames:v 1 -q:v 4 \u003Cframe>.jpg\n                Extract 2-3 frames per video at strategic timestamps:\n                  - t=2-5s : intro \u002F project name often appears\n                  - t=15-25s : middle \u002F detailed view of the tool\n                  - t=35-40s : outro \u002F CTA + URL often appears\n5. Vision    → Read each JPG — native multimodal sees:\n                  - Project names (top-left logos, browser tabs)\n                  - GitHub repo URLs (README banners, browser URL bar)\n                  - License badges (MIT \u002F Apache \u002F AGPL)\n                  - Star counts, version numbers\n                  - On-screen overlay text (Spanish captions, hashtags)\n6. Verify    → WebFetch or `gh search repos \"\u003Cname>\"` to confirm canonical URL exists\n                and license \u002F stars match what was on screen\n7. Cleanup   → rm -rf \u002Ftmp\u002F\u003Cprefix>_*  to release ~5MB per video\n```\n\n## Strategic timestamp picks\n\nMost tech-tip short-form videos follow this structure:\n\n| Timestamp | Typical content |\n|---|---|\n| 0-3s | Hook + caption with project pitch (\"AWS en local y gratis\") |\n| 3-10s | Project logo + landing page hero (\"Local AWS. Zero cost.\") |\n| 10-25s | Demo \u002F detail screen recording with on-screen features |\n| 25-35s | Stats \u002F license \u002F GitHub repo screenshot |\n| 35-45s | Outro CTA with URL + hashtags |\n\nSample frames at t=5, t=20, t=40 and you'll capture name, demo, and source URL in 90% of cases. Add t=12, t=32 for hard cases.\n\n## Cost profile (vs alternatives)\n\n| Approach | Tokens \u002F 1 video | API cost \u002F 1 video | Speed |\n|---|---|---|---|\n| WebFetch on TikTok URL | ~500 (bot-walled) | $0 | Fast, useless |\n| Browser scrape via agent-browser | ~5,000-10,000 | $0 | Slow, login-walled |\n| yt-dlp + OpenAI Whisper transcription | ~3,000 | ~$0.01 | Medium, needs API key |\n| **This skill (yt-dlp + frames + native vision)** | ~2,000 (3 frames × ~600 each) | **$0** | Fast, no API key |\n\nNative vision dominates for project-identification tasks. Transcription only wins when audio is the signal (which short-form tech video rarely is).\n\n## Permissions \u002F install prerequisites\n\n- `~\u002F.local\u002Fbin\u002Fyt-dlp` — install via standalone GitHub release (no PEP 668 conflict). Settings allowlist needs:\n  - `Bash(curl -fsSL https:\u002F\u002Fgithub.com\u002Fyt-dlp\u002Fyt-dlp\u002Freleases\u002F*)`\n  - `Bash(chmod +x \u002Fhome\u002F\u003Cuser>\u002F.local\u002Fbin\u002Fyt-dlp)`\n  - `Bash(~\u002F.local\u002Fbin\u002Fyt-dlp:*)`\n- `ffmpeg` — already present on most systems at `~\u002F.local\u002Fbin\u002Fffmpeg` or via `which ffmpeg`\n- `jq` — for parsing yt-dlp's JSON dumps\n- No external API keys required\n\n## Cleanup discipline\n\nAlways `rm -rf \u002Ftmp\u002F\u003Cprefix>_*` at the end of a mining run. Each video is ~5MB; a 10-video session = 50MB. Don't leave artifacts.\n\n## Limits \u002F when this fails\n\n- **Channel doesn't permit unauthenticated metadata access** — yt-dlp may need cookies; falls back to agent-browser\n- **Creator uses CapCut \u002F Premiere with minimal on-screen text** — name only spoken, never shown. Then audio transcription wins\n- **Branded text obscured by stylized fonts \u002F shaders** — vision misreads; use 2 frames to cross-check\n- **Project name is generic (\"API client\")** — vision identifies the screenshot but not the brand; WebFetch search step is essential\n\n## TikTok gotcha: do NOT add curl_cffi, and don't blindly upgrade yt-dlp\n\nVerified 2026-06-11 against a live TikTok pull. Two traps, both counter-intuitive:\n\n1. **curl_cffi impersonation BREAKS the TikTok extractor.** yt-dlp warns \"no impersonation target available\" on TikTok, which looks like the thing to fix. It is not. Inject `curl_cffi` and TikTok serves a challenge page with no rehydration JSON, so extraction dies with `ERROR: Unable to extract universal data for rehydration`. Plain urllib (no curl_cffi) gets the normal page and works. Leave that warning alone for TikTok.\n2. **yt-dlp versions after 2026.03.17 regressed TikTok** with the same rehydration error (2026.6.9 and master both fail, even without curl_cffi). Pin to a known-working build (`pipx runpip yt-dlp install \"yt-dlp==2026.3.17\"`) and only upgrade once an upstream TikTok fix lands. Test the actual target video after any yt-dlp change, never assume newer is safer.\n\nRecognition signal: `Unable to extract universal data for rehydration` on a TikTok URL = one of these two, not a network blip.\n\n## Anti-fabrication discipline (CRITICAL per CLAUDE.md)\n\nVision can hallucinate text it didn't actually see. **Every project name extracted by vision MUST be verified** via WebFetch or `gh search` before writing it into a brain skill or a client recommendation. The verification step is not optional. Workflow step 6 exists specifically to enforce this.\n\nIf verification fails (no repo found, name doesn't match search results), do NOT invent a URL or guess. Report \"vision saw X but no canonical repo found — manual verification needed\" and stop.\n\n## Related brain assets\n\n- `transcribe` — when audio IS the signal (interviews, lectures, podcasts)\n- `image-analyzer`, `gemini-vision` — for richer per-image analysis when frames need deeper interpretation\n- `agent-browser` — fallback when yt-dlp is blocked\n- `knowledge-corpus` — to organize what's mined into a queryable brain layer\n- Trend Researcher agent — natural consumer of mined social-video gems\n- Sister technique: the broader \"mine creator channels for OSS gems\" pattern that produced this skill, `claude-mem-persistent-memory`, `bruno-postman-alternative`, `coolify-self-hosted-paas`, `spacetimedb-backend-as-database`, and `vibevoice-elevenlabs-alternative` in a single research session\n",{"name":5,"description":7},"master","b4414608cd5a3f87d1525b518f1c6add8363540d","skills\u002Fsocial-video-mining",6,"https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato","https:\u002F\u002Fwww.dataqbs.com\u002Foctorato","https:\u002F\u002Fgithub.com\u002FCarlosCaPe","www.dataqbs.com"," AI-powered data engineering\r\n& consulting\r\n\r\nWe build data platforms that turn complexity into clarity — from cloud architecture and ETL pipelines to AI-driven","2026-11-14 00:21:00",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add CarlosCaPe\u002Foctorato --skill social-video-mining",{"tool":40,"command":41},"bunx","bunx skills add CarlosCaPe\u002Foctorato --skill social-video-mining",{"tool":43,"command":44},"pnpm","pnpm dlx skills add CarlosCaPe\u002Foctorato --skill social-video-mining","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fsocial-video-mining\u002FSKILL.md",true,[48],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"skill.json","REFERENCE","参考文档",334,"https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fsocial-video-mining\u002Fskill.json",[55,64,72,79,87,95],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":59,"licenseFileUrl":60,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M74EVAA4A4B4AKR5TT732GP4","yeet","Use only when the user explicitly asks to stage, commit, push, and open a GitHub pull request in one flow using the GitHub CLI (`gh`).","Apache-2.0","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fyeet\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fyeet\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":59,"licenseFileUrl":68,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M74EV4WMRDKP2NQS4YWR8PXH","wrangler","Cloudflare Workers CLI for deploying, developing, and managing Workers, KV, R2, D1, Vectorize, Hyperdrive, Workers AI, Containers, Queues, Workflows, Pipelines, and Secrets Store. Load before running wrangler commands to ensure correct syntax and best practices. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fwrangler\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fwrangler\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":19},"01M74EV05MHTXKEXD4B2QJEYTF","workspace-skill-discovery","Discover and include ALL skills across the workspace — not just global ones. Use at the start of every session or when the user asks about available skills. Ensures project-level skills (under .claude\u002Fskills\u002F or skills\u002F in any repo) are never missed.","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fworkspace-skill-discovery\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":59,"licenseFileUrl":83,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M74ETV78525B4VP2F28Z6NRD","workers-best-practices","Reviews and authors Cloudflare Workers code against production best practices. Load when writing new Workers, reviewing Worker code, configuring wrangler.jsonc, or checking for common Workers anti-patterns (streaming, floating promises, global state, secrets, bindings, observability). Biases towards retrieval from Cloudflare docs over pre-trained knowledge.","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fworkers-best-practices\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fworkers-best-practices\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":59,"licenseFileUrl":91,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M74ETR75NF67WGVEN7288ACV","winui-app","Create, build and design WinUI 3 desktop apps with C# and the Windows App SDK following Microsoft's official guidance and CommunityToolkit. Covers XAML, navigation, theming, accessibility and deployment.","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fwinui-app\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fwinui-app\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":19},"01M74ETJFRNPPNSQZ36C6HD6CW","windows-brain-script-portability","Scripts in ~\u002F.claude\u002Fscripts\u002F that run on a Windows brain: use the python3.cmd shim and not bare python3, and reconfigure stdout to UTF-8 if you print symbols. Use it when writing a brain script or triaging a Windows-only crash.","https:\u002F\u002Fgithub.com\u002FCarlosCaPe\u002Foctorato\u002Fblob\u002Fmaster\u002Fskills\u002Fwindows-brain-script-portability\u002FSKILL.md",[],[]]