The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external PDF documents, which represents an attack surface for indirect prompt injection. * Ingestion points: User-provided PDF files intended to be processed by the add_watermark_to_pdf tool. * Boundary markers: The instructions do not specify explicit delimiters or guardrails for the agent to ignore instructions that might be embedded within the PDFs being watermarked. * Capability inventory: The skill utilizes the add_watermark_to_pdf tool from the office-mcp server. It does not perform arbitrary file writes, network operations, or dynamic code execution beyond the scoped tool. * Sanitization: No specific sanitization or filtering of PDF content is mentioned before processing.

PDF Watermark