KY Markdown Rebuilder skill
A skill that rebuilds PDFs, decks and scans into page-aligned Markdown by checking extracted text against page images.
- 115 ★ on the repository
- Last commit
Minimal access
Why this level
- Document pages go to the agent's vision model, so confidential material should not be processed this way.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/ky-markdown-rebuilderDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- Document pages go to the agent's vision model, so confidential material should not be processed this way.
This is third-party code. Review the repository files before installing.
What it does
The skill converts visually complex documents into Markdown that can be checked against the original. Text is extracted first, pages are then rendered and content is verified against the visual structure. Scripts check page count, format, image links and the output contract; format drift, a lost page, a dead link or a broken table blocks delivery. Each page or slide gets its own section so neighboring pages do not mix. Headings, columns, cards, tables, diagrams and matrices are kept. Scans are first transcribed page by page by a vision model and only then fall back to Tesseract. Unreadable sums, units and small print are flagged explicitly, and the number of originals actually opened is recorded. Modes are `fast`, `page-aligned`, `deep-visual` and `transcribe`; the main output is `source.calibrated.md`.
Who it is for. Analysts and editors who need accurate Markdown from PDFs, decks and scans for further agent work.
Good fit when
- A PDF whitepaper with tables and diagrams loses order under ordinary conversion.
- Presentation slides are needed as separate sections with structure kept.
- A book scan is transcribed page by page without added commentary.
- A reading version of a long document is prepared after page-level calibration.
Not a fit when
- The text is plain or already Markdown, where an ordinary tool is faster.
- The document is classified and cannot go to a vision model.
- A legally exact transcription of handwritten documents is needed.
Example request
Rebuild `report.pdf` into page-aligned Markdown in `page-aligned` mode, flag unreadable sums and report how many pages were checked against the original.Limitations
Quality depends on the vision model and source resolution: small type and poor scans produce flags for unreadable text. External programs are needed, and the `preflight.py` check lists them: `pdftoppm`, `pdfinfo`, `soffice`, `markitdown`, `tesseract`. Document pages go to the agent's vision model, so confidential material should not be processed this way. No license is set in the repository. The `deep-visual` mode spends more time on page-level checks. The last commit dates from July 2026.
How to disable. Delete the skill folder from `~/.codex/skills` or `~/.claude/skills`. Markdown files and page images already created stay beside the source document.
FAQ
Which documents fit?
PDFs, PPT and PPTX decks, scans and page photos, Word-style reports, saved HTML pages, long screenshots and spreadsheets with dashboards.
How do the modes differ?
`fast` gives quick text conversion, `page-aligned` one section per page, `deep-visual` a full check against the original and `transcribe` verbatim transcription.
What about scans?
The skill transcribes them page by page with a vision model and uses Tesseract only as a fallback.
What file comes out?
`source.calibrated.md` by default; on request also `source.outline.md` and a separate translation file.
How do I update the skill?
Run `git pull --ff-only` in the skill folder under `~/.codex/skills` or `~/.claude/skills`.
Discussion
No questions or reviews yet.
Related
Skillsdoc-cleaner converts PDF, Office, Keynote, Numbers and EPUB into clean Markdown with intact tables and Big5 and CP950 encodings.
A Firecrawl skill and CLI that lets your agent convert Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV and PDF to Markdown on your own machine
A document pipeline skill: anything to markdown, markdown to clean html and publisher-grade docx, and html back to markdown
A pipeline turns PDF and EPUB textbooks into markdown and writes Obsidian notes with book citations and figures.
Microsoft's utility and MCP server that turn PDF, Word, Excel, PowerPoint, HTML and other files into Markdown for models
A skill and CLI that turn Markdown into DOCX, PDF, PPTX, XLSX, PNG and about a dozen more formats, all computed locally
README in short
The README explains the approach of text extraction, page rendering, original comparison and mechanical checks, lists input formats and modes and describes install and update.