The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/fill_fillable_fields.py implements a runtime monkey patch for the pypdf library. It dynamically replaces pypdf.generic.DictionaryObject.get_inherited with a custom implementation at execution time to work around a specific formatting bug in the library's handling of selection lists.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted PDF documents for text extraction, table analysis, and form filling. This creates a significant surface for indirect prompt injection where malicious instructions embedded in PDF content could influence the agent's behavior.
Ingestion points: Untrusted data enters the agent context through pypdf, pdfplumber, and pdf2image (via OCR) in scripts such as extract_form_field_info.py and convert_pdf_to_images.py.
Boundary markers: The instructions in forms.md and SKILL.md lack explicit boundary markers or warnings to the agent to treat extracted text as untrusted data.
Capability inventory: The skill possesses significant capabilities including file system write access (PdfWriter), complex PDF creation (reportlab), and command execution for external tools like qpdf and pdftotext.
Sanitization: There is no evidence of sanitization, filtering, or escaping of extracted PDF content before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The documentation (SKILL.md, reference.md) instructs the agent to install and use several third-party libraries including pypdf, pdfplumber, reportlab, pytesseract, pdf2image, pypdfium2, pdf-lib, and pdfjs-dist. While these are well-known industry standard packages, they represent an external dependency chain.

pdf