← ClaudeAtlas

markdown-section-summarizerlisted

Summarizes Markdown files section-by-section with strict sentence budgets derived from source section length. Use when you need deterministic per-section summary quotas, inventory-first workflows, and mechanical verification that each section has exactly floor(sqrt(n)) summary sentences.
msewell/agent-stuff · ★ 0 · Data & Documents · score 70
Install: claude install-skill msewell/agent-stuff
# Markdown Section Summarizer Create section-aware Markdown summaries with strict, auditable sentence counts. ## Workflow 1. Build an inventory from the source Markdown, specifying the document language. 2. Generate the summary file directly from the inventory. It contains all section headings, preserved image embeds, and per-section sentence-count comment blocks. 3. Fill in each section's content by writing the target number of sentences after each comment block. 4. Verify that the summary exactly matches the inventory sentence targets. 5. If verification fails, revise the flagged sections and re-run verification. ## Setup ```bash python3 -m pip install -r markdown-section-summarizer/scripts/requirements.txt ``` ## Commands ```bash # Step 1 — build inventory (--language is required; use the ISO 639-1 code of the source document) python3 markdown-section-summarizer/scripts/build_inventory.py input.md --language en -o input.inventory.json # Step 2 — generate summary file (headings + preserved images + sentence-count placeholders) python3 markdown-section-summarizer/scripts/generate_summary.py input.inventory.json -o input.summary.md # Step 3 — fill in input.summary.md: write the target number of sentences after each comment block # Step 4 — verify python3 markdown-section-summarizer/scripts/verify_summary.py input.md --inventory input.inventory.json --summary input.summary.md ``` Supported language codes (pysbd): `am ar bg da de el en es fa fr hi hy it ja kk m