The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run in scripts/to_pdf.py to execute pandoc for document format conversion and fc-cache for font management. It also launches a bundled helper script _render.py via the Python interpreter to finalize PDF rendering.
[EXTERNAL_DOWNLOADS]: The scripts/emoji_img.py script fetches SVG assets from the jsDelivr CDN (cdn.jsdelivr.net) to support emoji rendering. These downloads are directed to the official Twitter Twemoji repository and are used solely as static image data.
[PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection as it processes untrusted Markdown and HTML data.
Ingestion points: User-provided content is read from the filesystem into scripts/to_pdf.py and scripts/emoji_img.py.
Boundary markers: No explicit delimiters or instructions are used to isolate user-controlled file content from the agent's instructions during processing.
Capability inventory: The skill has access to shell execution (subprocess.run) and network requests (urllib.request).
Sanitization: Plain text is sanitized using html.escape, and Markdown is converted via Pandoc; however, complex HTML inputs are rendered without additional sanitization for embedded prompt instructions.

pdf-converter