Overview
Requires Python 3.11+, uv, ffmpeg, and ffprobe on PATH. Tesseract is governed by the selected OCR policy. Scripts use exact pinned PEP 723 dependencies via `uv run`.
Operations
Each script returns structured JSON — see the script contract.
| Op ID | Script | Common flags |
|---|---|---|
program_master.analyze_structure | analyze_structure.py | --input, --fps |
program_master.compile | compile.py | — |
program_master.compile_pdf | compile_pdf.py | --output |
program_master.detect_black_silence | detect_black_silence.py | --input |
program_master.detect_blacks | detect_blacks.py | --input |
program_master.detect_silence | detect_silence.py | --input |
program_master.doctor | doctor.py | — |
program_master.label_segments | label_segments.py | --input, --timecode, --fps |
program_master.run_report | run_report.py | --input, --output, --timecode |
program_master.schema | schema.py | — |
program_master.self_test | self_test.py | --input, --output |
program_master.validate_report | validate_report.py | --output |
Do not use for
- hard cuts or camera shots inside content (use `shots`)
- subtitle or caption transcription (use `forced-narrative-exact` or `vision-analysis`)
- SCTE-only segmentation without black+silent structural evidence
- Related skills: `inspect`, `timecode`, `forced-narrative-exact`, `vision-analysis`, and `shots`.