The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes external binaries to perform document processing tasks. It uses subprocess.run with list-based arguments to invoke pandoc for Markdown-to-HTML conversion, Google Chrome for headless PDF rendering, and Poppler utilities (pdftoppm, pdfinfo, pdftotext) for preview generation and typography linting. This implementation follows security best practices to avoid shell injection.
[INDIRECT_PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection as it processes user-supplied markdown files and instructs the agent to visually verify the converted output.
Ingestion points: Untrusted data enters the environment through input markdown files processed by scripts/md_to_pdf.py and scripts/batch_convert.py.
Boundary markers: There are no explicit delimiters or specific instructions for the agent to disregard commands or directives that might be embedded within the documents being processed.
Capability inventory: The skill has the capability to read and write files and execute shell commands via orchestrating external tools.
Sanitization: The skill does not filter or sanitize the contents of the markdown files before they are passed to the conversion pipeline.

pdf-creator