The Agent Skills Directory

[PROMPT_INJECTION]: The skill extracts text and metadata from PDF files and ingests them into the agent's context. This creates a surface for indirect prompt injection, where a maliciously crafted PDF could contain instructions that manipulate the agent's behavior.
Ingestion points: SKILL.md text extraction examples, scripts/extract_form_field_info.py, and scripts/extract_form_structure.py.
Boundary markers: No explicit delimiters or instructions are used to distinguish between data and potential instructions within the PDF content.
Capability inventory: The agent has access to file system writes and shell execution of PDF utilities.
Sanitization: There is no evidence of text sanitization or filtering before the data is processed.
[COMMAND_EXECUTION]: The skill relies on shell-based execution for tools like qpdf, pdftotext, pdftk, and magick. This presents a command injection risk if filenames or parameters are derived from untrusted user input without adequate sanitization.

pdf