The Agent Skills Directory

COMMAND_EXECUTION (LOW): The script scripts/compile.py invokes the typst and latexmk binaries using subprocess.run. While the implementation follows best practices by using list-based arguments and avoiding shell=True, the execution of document compilation engines on content generated from untrusted sources presents a minor security boundary risk inherent to these tools.
PROMPT_INJECTION (LOW): The skill is susceptible to indirect prompt injection through its paper conversion functionality.
Ingestion points: External data is ingested through the from-paper workflow as defined in WORKFLOWS.md, which processes content from academic papers.
Boundary markers: Absent. The provided documentation and resource files do not specify the use of delimiters or 'ignore' instructions for the ingested text.
Capability inventory: The skill possesses the ability to execute system commands (typst, latexmk) and perform recursive file system operations across the project root.
Sanitization: Absent. No sanitization logic for input text is present in the provided Python scripts or resource-based instructions.

academic-slides