The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a validation script (scripts/validate_format.py) that executes local system utilities pdfinfo and pdffonts via the subprocess module to check document compliance. These calls use argument lists rather than shell strings, which follows security best practices for avoiding shell injection.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data, including LaTeX templates and PDF files for customization and validation. While this creates a potential surface for indirect prompt injection if the agent parses malicious content within these files, the skill's operations are restricted to specific formatting tasks, and no evidence of exploitation was found.
Ingestion points: scripts/customize_template.py (LaTeX templates) and scripts/validate_format.py (PDF files).
Boundary markers: Not explicitly implemented in the scripts' text processing logic.
Capability inventory: The skill has file read/write permissions and can execute subprocesses for PDF analysis.
Sanitization: Basic regex-based text replacement is used in customization; path strings are used for subprocess calls.

venue-templates