Update to YouTube Agent

{
  "description": "Turn a raw recording's transcript into an edit decision list - dead air, filler cues and retakes, with timecodes. Use for \"edit this\", \"cut the dead space\", \"tighten this video\", \"I rambled\", or any request to shorten footage from a transcript.",
  "included_files": [
    {
      "relative_path": "deadair.py",
      "size_in_bytes": 4244
    }
  ],
  "name": "yt-edit",
  "skill_md_contents": "---\nname: yt-edit\ndescription: >-\n  Turn a raw recording's transcript into an edit decision list - dead air,\n  filler cues and retakes, with timecodes. Use for \"edit this\", \"cut the\n  dead space\", \"tighten this video\", \"I rambled\", or any request to shorten\n  footage from a transcript.\n---\n\n# yt-edit\n\nUse scripts relative to this skill directory, not the user’s working directory. If local execution or inputs are unavailable, explain the limitation and provide a clearly labeled manual alternative; do not claim a tool ran. Scores and thresholds are heuristics, not guarantees or causal proof. Use only accessible or user-provided data; never invent analytics.\n\nAn edit decision list from a timestamped transcript. It prints the cuts. You apply them.\n\n```bash\npython3 deadair.py transcript.srt              # srt, vtt or whisper json\npython3 deadair.py transcript.srt --floor 0.35 --json\n```\n\nNo transcript yet? Ask for one, or produce one first - `whisper`, `faster-whisper`, or the caption\ntrack YouTube generates on an unlisted upload all work. Do not guess at timings.\n\n## What it finds\n\n- **DEAD** - gaps longer than the floor, trimmed from the MIDDLE so both sides keep a breath.\n  Cutting flush against speech is what makes a tightened take sound gasping.\n- **FILLER** - cues that are nothing but \"um\", \"so yeah\", \"basically\".\n- **REPEAT** - a sentence restarted. Compared against the last cue that was actually speech, not\n  the literal previous cue, because most retakes have an \"um\" between the two attempts.\n\n## What it will not do\n\nIt does not touch media. It has no opinion about your B-roll. A 40% cut on the report is a 40% cut\nof SPEECH, and if the video has a long silent demo in it that number is wrong - check the report\nagainst the footage before you trust the runtime at the bottom.\n\n## The gate\n\nNothing here publishes. This skill writes and you publish. Every output ends in a block the user\ncopies, and the last line of every run is the question: **ship it, or change it?**\n"
}