The Agent Skills Directory

[COMMAND_EXECUTION]: The script executes the mmdc (Mermaid CLI) binary to render diagrams. This execution is performed using a list of arguments without invoking a shell, which prevents shell command injection vulnerabilities. Evidence: subprocess.run(cmd, ...) in scripts/converter.py uses a list-based command structure.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Markdown content as its primary input. While this represents a data ingestion surface, the risk is limited. Evidence Chain:
Ingestion points: The input file path in scripts/converter.py reads user-provided Markdown files.
Boundary markers: None present in instructions or script.
Capability inventory: subprocess.run for diagram rendering, file read/write for PDF generation.
Sanitization: Employs the markdown library and WeasyPrint, which do not execute JavaScript or arbitrary scripts from input content.
[DYNAMIC_EXECUTION]: The skill dynamically assembles HTML from Markdown, templates, and CSS at runtime to generate the PDF output. This is a core part of its documented functionality and is handled through established libraries without the use of dangerous functions like eval() or exec() on untrusted data.

convert-md-to-pdf