Skip to content

Latest commit

 

History

History
29 lines (22 loc) · 1.48 KB

File metadata and controls

29 lines (22 loc) · 1.48 KB
description Generate a cut-for-time report from prepared transcript artifacts

Generate a cut-for-time report for: $@

Workflow:

  1. Identify the target episode slug and the matching transcript directory or prepared transcript artifacts.
  2. If dist/analysis/<slug>/transcript_chunks.md or dist/analysis/<slug>/transcript_index.json do not exist yet, run: uv run python scripts/prepare_transcript_analysis.py <transcript-dir> --output-dir dist/analysis/<slug> --slug <slug> --plain-output-names
  3. Read the prepared transcript artifacts. For long transcripts, use the chunk index first and then inspect the weakest or most repetitive chunk sections.
  4. Use pi-driven judgment to identify sections that can be trimmed or cut.
  5. Write the final report to dist/analysis/<slug>/cut_report.md.
  6. Reply with the output path and a short summary of the biggest cut opportunities.

Output expectations:

  • include exact time ranges for each suggested cut or trim
  • explain why each section is weak, repetitive, speculative, low-energy, or too niche
  • estimate likely time savings when possible
  • distinguish between hard cuts and lighter trims when useful
  • mention especially strong sections that should definitely stay if that helps the edit

Quality bar:

  • optimize for a tighter, stronger episode rather than maximizing total cuts
  • avoid cutting sections only because they are technical if they still have clear payoff
  • prefer cuts that reduce repetition, dead setup, or low-signal wandering