The Agent Skills Directory

[SAFE]: The skill uses established, well-known libraries (such as pypdf, pdfplumber, and reportlab) for its stated purpose of PDF processing. All provided scripts are functional and do not contain malicious code or hidden operations.
[DYNAMIC_EXECUTION]: The script fill_fillable_fields.py performs a runtime monkeypatch of the pypdf library's get_inherited method. This is used to resolve specific issues with how form field options are retrieved and is a legitimate implementation detail rather than a security risk.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted PDF data. While this creates a potential surface for indirect prompt injection (where a PDF could contain instructions for the agent), the skill's capabilities are focused on document transformation and it does not pass extracted text to dangerous sinks like eval() or shell execution, keeping the risk level low and typical for document processing tools.
[COMMAND_EXECUTION]: The documentation provides instructions for using standard CLI tools like qpdf, pdftotext, and magick for image processing. These are common utilities for the intended tasks and do not involve unsanitized user input in a way that would lead to command injection.

pdf