Warn
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The script
scripts/fill_fillable_fields.pyimplements a monkeypatching pattern to modify the behavior of thepypdflibrary at runtime. Specifically, the functionmonkeypatch_pydpf_methodredefinespypdf.generic.DictionaryObject.get_inheritedto handle a bug in how selection list fields are processed. While intended as a maintenance fix, this dynamic modification of library bytecode at runtime is a sophisticated execution pattern. - [INDIRECT_PROMPT_INJECTION]: The skill functions by ingesting and processing external PDF files provided by users or third parties, which may contain malicious instructions hidden in text or metadata.
- Ingestion points: PDF files are read and analyzed by
scripts/check_fillable_fields.py,scripts/extract_form_field_info.py,scripts/fill_fillable_fields.py, andscripts/fill_pdf_form_with_annotations.py. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions to prevent the agent from obeying commands embedded within the PDF content.
- Capability inventory: The skill has significant capabilities including file read/write access for PDF and PNG creation, and command execution for PDF processing tools, allowing ingested instructions to potentially influence downstream actions.
- Sanitization: There is no evidence of sanitization, filtering, or escaping of extracted text before it is used to generate JSON metadata or annotations.
Audit Metadata