The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user-provided PDF documents, which could contain malicious instructions aimed at overriding agent behavior or exfiltrating data using the skill's file-writing capabilities. * Ingestion points: The skill reads PDF content using the PyMuPDF library as described in Step 2 of SKILL.md. * Boundary markers: The skill does not define specific delimiters or instructions to ignore potential commands embedded within the extracted PDF text. * Capability inventory: The skill has access to Bash (for command execution), Read, and Write/Edit tools (for filesystem operations). * Sanitization: There is no evidence of sanitization, validation, or filtering of the extracted text before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill utilizes runtime script generation to handle PDF text extraction. * Evidence: In Step 2 of SKILL.md, a Python script is dynamically constructed and executed via the Bash tool to parse the provided PDF files. * Risk: While the script follows a static template, the practice of generating and executing code at runtime is a dynamic execution pattern that increases the attack surface.

epd-parser