The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/setup.sh script executes commands with administrative privileges (sudo apt-get install) to install system dependencies, which is a high-risk operation for an agent skill.
[COMMAND_EXECUTION]: The scripts/md_to_pdf.py script uses subprocess.run to execute various external binaries (mmdc, pandoc, node) to process user-provided markdown content.
[EXTERNAL_DOWNLOADS]: The setup script performs global installations of packages from external registries using npm install -g and pip install, and triggers the download of the Chromium browser binary.
[PROMPT_INJECTION]: The skill processes untrusted markdown data through multiple rendering stages (Mermaid, KaTeX, Playwright/Chromium), presenting a significant surface for indirect prompt injection.
Ingestion points: The main script scripts/md_to_pdf.py reads content directly from the user-provided markdown file.
Boundary markers: No security delimiters or warnings are used to prevent the downstream rendering engines from interpreting malicious instructions embedded in the markdown.
Capability inventory: The conversion process involves file system writes, external command execution, and a headless browser capable of local file access and network requests.
Sanitization: The skill lacks validation or sanitization of the input markdown, Mermaid diagrams, or LaTeX equations before they are passed to external processors.

md-to-pdf