Docs Converting PDF To Markdown

Converting PDF to Markdown

A Markdown file converted from a PDF must be a verbatim, complete representation of its source: every passage, table, heading level, list-nesting depth, figure, and diagram preserved.

When This Skill Loads

Auto-loads for pdf-to-md-maker, pdf-to-md-checker, and pdf-to-md-fixer.

Lifecycle Delegation

Quality-gate invocations may pass exact delegated-gate-ids under Lifecycle Validation Ownership. Checker and fixer omit only matching generic Markdown predicates. PDF/source fidelity remains authoritative. Accept lifecycle-evidence: checkers preserve it; fixers scope-intersect changed files and return updated-lifecycle-evidence. Omitted delegation preserves standalone full behaviour.

Reference Modules

• checking-fidelity-criticality-and-format.md and checking-fidelity-workflow.md — the checker's seven-dimension validation workflow, crane check-* commands, criticality table, and audit report format.
• making-conversions-detect-and-extract.md and making-conversions-assemble-and-write.md — the maker's PDF-type detection, text/OCR extraction, chunk-to-Markdown conversion, assembly, and output steps.
• fixing-conversions-confidence-and-priority.md and fixing-conversions-operations-and-report.md — the fixer's confidence assessment, priority order, per-finding fix operations, and fix report format.

Tools Usage

• Bash: crane (pdf/text/heading/nesting/table/figure/mermaid/ocr/report/skiplist commands), diff for comparison
• Read: source Markdown and temporary extracted text files
• Glob: locate the MD file when path not specified
• Grep: search MD for text segments, count figures, find Mermaid blocks
• Write: progressive audit/fix reports to local-tmp/pdf-to-md/

Reference Documentation

• pdf-to-md-quality-gate workflow