Use when cleaning Taiwan tax-law course transcripts or rough notes into structured Markdown. Adds context-aware headings, optimizes heading hierarchy, and merges fragmented short lines into readable paragraphs, with a default target of about 6 short clauses per line.
Use this skill for raw class transcripts under mkdocs/My_Notes/** when the file is fragmented into many short lines and lacks usable section structure.
Turn a raw transcript into a note that:
### and ### headings based on the lecture flow## for major topic shifts and only when the subsection is clearly nested under the current section.###借名行為的民事法評價討論一 or 補充#: one file-level lecture title##: main lecture blocks###: doctrinal tests, tax-item comparisons, or issue-specific subparts# directly to ##### heading with no leading blank line, but it should still have exactly one blank line after itUse scripts/merge_short_lines.py for deterministic paragraph cleanup before or after heading insertion.
Example:
python3 .codex/skills/tax_transcript_polisher/scripts/merge_short_lines.py \
--in 'mkdocs/My_Notes/1142/課_四89_所得稅法專題研究/逐字稿初稿/W04_0319.md' \
--out 'mkdocs/My_Notes/1142/課_四89_所得稅法專題研究/逐字稿初稿/W04_0319.md'