interview-stringout

I'd like you to add a skill to Wideframe called "interview-stringout".

What it should do:
Transcribe a batch and lay the spoken takes into a sequence, ready for Premiere.

Before you run anything, ask me to confirm:
- Radio edit (remove filler / false starts) or verbatim?
- Minimum clip length + silence-gap split threshold (seconds)
- Which speaker(s) to include
- Sequence frame rate + resolution

Then do the work:
1. caltools media-transcript-get-batch over $1 to get word-level timestamps for every file.
2. Segment the transcript into spoken runs using the confirmed silence-gap threshold.
3. If radio edit: mark filler tokens (um, uh, like) and false-start spans for removal; keep a log.
4. caltools sequence-create with the confirmed fps/resolution.
5. caltools sequence-clips-add each kept segment in source order; sequence-clips-update to trim handles.
6. Export with the native .prproj exporter.

Do not grade your own work. When the output is ready, hand it to a separate verification subagent that has its own context window. Give it the output, the choices I confirmed, and the pass criteria below. It reviews the quality, writes specific feedback, and returns PASS or FAIL.

Pass criteria:
- Each kept spoken segment appears exactly once, in order.
- A removed-segments log lists every cut span with timecodes.
- The exported sequence has zero offline/unlinked clips.

If it returns FAIL, apply the feedback, redo the affected work, and verify again with a new subagent. Repeat until it returns PASS, then show me the result.

Once it works, save it as a reusable skill named "interview-stringout" so I can run it again by saying /interview-stringout.