generate-pdf Skill | Agent Skills

Generate a production-grade branded PDF from a markdown source file using the pdf-factory pipeline.

$ARGUMENTS

Parse arguments: <markdown-file> [--brand <name>] [--output <path>]

<markdown-file> -- path to the source markdown file (required)
--brand <name> -- brand kit name: bluewaves, wave-artisans, or decathlon (default: bluewaves)
--output <path> -- output PDF path (default: same directory as input, .pdf extension)

Steps

Parse arguments from $ARGUMENTS. Extract markdown file path, brand name, and output path.

Verify dependencies are installed:

python3 -c "import xhtml2pdf, reportlab, pypdf, markdown, lxml, PIL, html5lib, cssselect2" 2>/dev/null

If this fails, run the install-docs-deps skill first.

Resolve brand kit at ${SKILL_ROOT}/../brand-<name>/. Verify the directory exists and contains assets/manifest.json.

Read the markdown file and extract frontmatter metadata (title, subtitle, author, date). Convert markdown to HTML:

python3 -c "
import markdown, json, sys, re
with open('<markdown-file>') as f:
    source = f.read()
# Extract YAML frontmatter
fm_match = re.match(r'^---\n(.*?)\n---\n', source, re.DOTALL)
meta = {}
if fm_match:
    for line in fm_match.group(1).splitlines():
        if ':' in line:
            key, val = line.split(':', 1)
            meta[key.strip()] = val.strip()
    source = source[fm_match.end():]
# Convert to HTML
html = markdown.markdown(source, extensions=['tables', 'fenced_code', 'codehilite', 'toc', 'meta', 'attr_list'])
with open('<output-dir>/content.html', 'w') as f:
    f.write(html)
with open('<output-dir>/metadata.json', 'w') as f:
    json.dump(meta, f, indent=2)
print('Parsed markdown and extracted metadata.')
"

Render content pages:

python3 ${SKILL_ROOT}/../pdf-factory/scripts/render.py \
  --brand ${SKILL_ROOT}/../brand-<name> \
  --input <output-dir>/content.html \
  --output <output-dir>/content-pages.pdf

Compose final document:

python3 ${SKILL_ROOT}/../pdf-factory/scripts/compose.py \
  --brand ${SKILL_ROOT}/../brand-<name> \
  --content <output-dir>/content-pages.pdf \
  --metadata <output-dir>/metadata.json \
  --output <output-path>

Validate output:

python3 ${SKILL_ROOT}/../pdf-factory/scripts/validate_output.py \
  <output-path> \
  --brand ${SKILL_ROOT}/../brand-<name>

Clean up temporary files (content.html, content-pages.pdf, metadata.json).
Report results -- output path, page count, file size, and validation status.

Agent Skills: generate-pdf

Install this agent skill to your local

Skill Files

Steps