メニュー
Fetch latest videos from YouTube channels, extract transcripts, and generate a digest with key insights. Use when asked for YouTube digest, channel updates, or video summaries.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Manage the Trinity AGI auth service and RBAC system -- database schema, role hierarchy, permission management, JWT flow, audit logging, and API endpoints.
Sync files between the host machine and the running Trinity Docker containers — config, agent files, skills, and extensions.
Develop the Trinity AGI Flutter Web Shell -- understand the architecture, A2UI renderer, WebSocket models, dual-client pattern, design tokens, and build workflow.
Operate the Trinity AGI Kubernetes infrastructure -- Helm charts, multi-tenant per-user OpenClaw pods, Vault Agent Injector secrets, minikube dev, and production deployment.
Operate the OpenClaw Gateway that powers Trinity AGI — configure providers, manage sessions, use the CLI, and understand the WebSocket protocol.
Manage the Trinity AGI Docker Compose stack — start, stop, logs, rebuild frontend, and full deploy.
| name | youtube-digest |
| description | Fetch latest videos from YouTube channels, extract transcripts, and generate a digest with key insights. Use when asked for YouTube digest, channel updates, or video summaries. |
| homepage | https://github.com/hesamsheikh/awesome-openclaw-usecases |
| metadata | {"openclaw":{"emoji":"📺","requires":{"bins":["uv"]},"primaryEnv":"YOUTUBE_API_KEY","install":[{"id":"uv-brew","kind":"brew","formula":"uv","bins":["uv"],"label":"Install uv (brew)"}]}} |
Fetch latest videos from YouTube channels, extract transcripts, and produce a digest.
uv run {baseDir}/scripts/digest.py --channels "@Fireship,@lexfridman,@ThePrimeTimeagen" --hours 48
With transcripts and summaries:
uv run {baseDir}/scripts/digest.py --channels "@TED,@Fireship" --hours 48 --transcript --max-videos 10
Keyword search mode:
uv run {baseDir}/scripts/digest.py --search "OpenClaw,AI agents" --hours 72 --transcript
JSON output:
uv run {baseDir}/scripts/digest.py --channels "@Fireship" --output json
@handle to channel ID via YouTube Data API v3youtube-transcript-api (free, no API key, no binary)--seen-file to avoid reprocessinguv run {baseDir}/scripts/digest.py --channels "@Fireship" --seen-file seen-videos.txt
Only new (unseen) videos are processed. After processing, video IDs are appended to the file.
Every morning at 8am, fetch latest videos from @TED, @Fireship, @lexfridman and give me a digest with key insights.
Save my channel list to memory so I can add/remove channels later.
YOUTUBE_API_KEY — YouTube Data API v3 key (required for channel resolution and upload listing; get one at https://console.cloud.google.com/)youtube-transcript-api.channel/latest listing requires YOUTUBE_API_KEY.yyyy-mm-dd-yt-digest.md.