The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes external PDF files which are untrusted data sources. This creates a surface for indirect prompt injection (Category 8) where instructions hidden in the document could attempt to influence the agent's behavior.
Ingestion points: Data is read from PDFs in scripts/extract_form_field_info.py, scripts/check_fillable_fields.py, and scripts/fill_fillable_fields.py.
Boundary markers: No explicit instructions or delimiters are provided to the agent to disregard instructions found within the processed PDF content.
Capability inventory: The skill provides capabilities for local file system read/write operations and execution of PDF manipulation scripts.
Sanitization: The skill does not perform content sanitization on the text or metadata extracted from PDFs before it is presented to the agent.
[COMMAND_EXECUTION]: The skill references several Python-based utilities and CLI tools (qpdf, pdftk) for PDF manipulation. It also includes a monkeypatching implementation in scripts/fill_fillable_fields.py to modify the pypdf library at runtime. While these tools are used for legitimate PDF processing tasks, they provide significant local system interaction capabilities.

pdf