Subtitle Packager

Follow shared release-shell rules in:

• postplus-shared release-shell rules

Use this skill when the transcript already exists and the next problem is:

• chunking a normalized transcript into subtitle-sized units
• ASS generation
• SRT generation
• VTT generation
• readable caption chunking
• packaging timed transcript data for editors

This skill should not call STT models itself.

Input Rule

Expect:

• a normalized-transcript.json file produced by video-transcription

If timing is missing, say so plainly.

Do not fake subtitle timing unless the user explicitly asks for heuristic timing.

Scripts

• scripts/chunk_normalized_transcript.mjs
• scripts/render_ass_from_normalized.mjs
• scripts/transcript_json_to_srt.mjs
• scripts/text_to_srt.mjs

Read These References

• references/ass-contract.md
• references/chunk-modes.md
• references/output-shape.md

Release-Shell Execution Contract

• keep chunking inputs, intermediate subtitle JSON, and render artifacts under <work-folder>/.postplus/subtitle-packager/
• keep only final user-facing subtitle exports outside .postplus/
• start with a bounded first pass on one normalized transcript before broader batch packaging
• if normalized timing is missing or invalid, stop immediately instead of inventing subtitle timing or switching to ad hoc shell glue