Podcast Intel
从配置的 RSS 播客源抓取新单集，经音频转写、主题分段、结合用户兴趣打分后生成排序的“值得收听”推荐，支持 JSON/Markdown/TTS 输出，并维护本地收听日记；提供简报、单集分析、日记查看和快速推荐四类工具。
安装技能
复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。
执行命令安装此 skill： npx secure-skills-cli add podcast-intelMetadata
- name
- podcast-intel
- description
- Podcast intelligence engine. Transcribes, segments, summarizes, and scores podcast episodes from RSS feeds. Generates "worth your time" recommendations with cross-source deduplication and maintains a local consumption diary. Use this skill when the user asks about podcasts, wants podcast recommendations, says "what podcasts should I listen to", "podcast briefing", "what's worth listening to", "podcast summary", "summarize this podcast", "podcast diary", "what did I listen to", or any request involving podcast content analysis. Also activates for: "brief me on podcasts", "podcast recap", "new episodes", "what's new from [podcast name]", "is this episode worth it".
- metadata
- {"openclaw":{"emoji":"🎧","primaryEnv":"OPENAI_API_KEY","requires":{"bins":["python3","ffmpeg"],"env":["OPENAI_API_KEY"],"optionalEnv":["WHISPER_MODEL","PODCAST_INTEL_FEEDS","PODCAST_INTEL_INTERESTS"]}}}
- tools
- [{"name":"podcast_briefing","description":"Full intelligence pipeline: fetch new episodes from configured feeds, transcribe audio, segment by topic, analyze against user interests, and generate ranked \"worth your time\" recommendations. Appends to local consumption diary.\n","command":"cd {baseDir} && python3 scripts/main.py","args":[{"name":"hours","type":"number","description":"Look back this many hours for new episodes (default: 168 = 7 days)","default":168},{"name":"top","type":"number","description":"Return only the top N ranked recommendations (default: 10)","default":10},{"name":"output","type":"string","description":"Output format: json, markdown, or tts (default: markdown)","default":"markdown","options":["json","markdown","tts"]},{"name":"dry_run","type":"boolean","description":"Analyze episodes but don't write to diary","default":false}],"category":"podcast","requiredEnv":["OPENAI_API_KEY"]},{"name":"podcast_analyze","description":"Analyze a specific podcast episode by URL. Transcribes, segments, scores relevance and novelty against your interests.\n","command":"cd {baseDir} && python3 scripts/main.py","args":[{"name":"episode_url","type":"string","description":"Direct URL to podcast audio file or RSS feed entry URL","required":true},{"name":"show_name","type":"string","description":"Name of the podcast show (optional, for context)"},{"name":"output","type":"string","description":"Output format: json or markdown","default":"markdown","options":["json","markdown"]}],"category":"podcast","requiredEnv":["OPENAI_API_KEY"]},{"name":"podcast_diary","description":"Display your podcast consumption diary — a structured log of episodes you've been briefed on, topics covered, and recommendations.\n","command":"cd {baseDir} && python3 scripts/diary.py","args":[{"name":"show","type":"boolean","description":"Display diary entries","default":true},{"name":"last","type":"number","description":"Show entries from the last N days (default: 7)","default":7},{"name":"format","type":"string","description":"Output format: json or markdown","default":"markdown","options":["json","markdown"]}],"category":"podcast"},{"name":"podcast_recommend","description":"Quick recommendations from already-transcribed episodes. No new transcription — uses cached transcripts and instant scoring.\n","command":"cd {baseDir} && python3 scripts/main.py","args":[{"name":"top","type":"number","description":"Return only the top N ranked recommendations","default":5},{"name":"recommend_only","type":"boolean","description":"Skip transcription, use cache only","default":true},{"name":"output","type":"string","description":"Output format: json or markdown","default":"markdown","options":["json","markdown"]}],"category":"podcast","requiredEnv":["OPENAI_API_KEY"]}]