| name | md-to-pdf |
| description | Render a Markdown file in this project to a styled PDF (with embedded images). Use when the user asks to "render/convert/export a .md to PDF", "rebuild/regenerate the PDF", or mentions creating the Berufsvorstellung / Diplom Konzept PDF. |
Markdown → PDF
Renders a Markdown file in this project to a child-friendly, styled A4 PDF with
all images embedded. There is no pandoc / wkhtmltopdf / node here — this uses
Python's markdown library plus Chrome (or Edge) headless.
How to run
python ".claude\skills\md-to-pdf\md_to_pdf.py" "Berufsvorstellung Systems Engineer.md"
- One argument: the
.md file. The PDF lands next to it (same name, .pdf).
- Optional second argument: a custom output path.
python -m pip install markdown is auto-run if the library is missing.
The script is self-contained: it builds styled HTML, inlines every local image
as a base64 data URI, renders with a headless browser into a temp folder, then
copies the finished PDF into place.
Environment quirks the script already handles
- Local images don't show up unless inlined — Chrome headless blocks
file:// images. The script base64-embeds them into the HTML.
- Chrome can't write into the SynologyDrive folder ("Zugriff verweigert").
The script renders to a temp dir and copies the PDF in afterwards.
- A running/stale Chrome instance hijacks the render and produces an
image-less PDF. The script forces a throwaway
--user-data-dir.
Sanity checks
- The full document (with all images) is several hundred KB. An image-less
render is only ~40 KB — if you get that, something regressed.
- Count pages:
python -c "import re; d=open('FILE.pdf','rb').read(); print(len(re.findall(rb'/Type\s*/Page[^s]', d)))"
Styling
Defined in the CSS constant in md_to_pdf.py: blue headings, yellow
highlighted > **Frage:** blockquote boxes, bordered tables, centered images.
Adjust there if the look needs to change.
