The Agent Skills Directory

[COMMAND_EXECUTION]: The skill directs the agent to execute local Python scripts via the Bash tool, specifically scripts/generate_schematic.py for figure creation and skills/scientific-slides/scripts/pdf_to_images.py for converting slide decks into images during the review process.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because its operational workflow requires the ingestion and analysis of untrusted scientific documents and presentations.\n
Ingestion points: Untrusted data enters the agent's context through document file reading and the sequential inspection of slide images converted from PDFs.\n
Boundary markers: The instructions do not define delimiters or provide specific guidance to the agent to ignore or isolate instructions that might be embedded within the research documents being reviewed.\n
Capability inventory: The skill is configured with the Read, Write, Edit, and Bash tools, representing a broad set of capabilities that could be misused if the agent were to follow malicious instructions from a processed document.\n
Sanitization: There is no evidence of input validation, filtering, or sanitization for the content extracted from external manuscripts or slide decks.

peer-review