Use this skill when the user needs a gold-reference extraction/transcript, consensus evaluation, or provider comparison report from a multi-provider AutoShow image, music, OCR, STT, text/write, TTS, URL, or video run. Apply it to build category-specific consensus packets, generate reports, and compare providers with category-appropriate ranking and tier summaries.
Normalize AutoShow episode script Markdown without changing story content. Use when Codex needs to clean, format, or validate AutoShow script files, especially Markdown under input/episode-scripts/**, by enforcing the project script shell, speaker/delivery spacing, parenthetical formatting, line whitespace, and final-newline rules.
Use this skill when the user wants an ASR, meeting, interview, podcast, video, call, or voice-note transcript cleaned without changing meaning. Apply it for duplicate-word removal, filler cleanup, punctuation, grammar and spelling repair, timestamp or speaker-label preservation, and reconciliation against alternate provider transcripts -- even if the user simply says "clean this up" or "fix this text" and the input is clearly a transcript.
Clean raw OCR output into the most accurate, faithful text representation of a source document for archiving, reference, and search (not narration). Use when working on OCR-derived exports, scanned-book pages, or extraction.txt/chapter files that need artifact repair, hyphenation rejoining, structure preserved as lightly-structured Markdown, and footnotes/citations kept intact while preserving original wording.
Use this skill when the user wants long-form plaintext prepared for natural, service-agnostic text-to-speech narration. Apply it to OCR-derived EPUB exports, book chapters, articles, or chapter text files that need artifact removal, paragraph repair, punctuation cleanup, footnote removal, boundary continuity checks, and pronunciation-friendly edits while preserving meaning.