The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted PDF files to extract text and form structures, creating an attack surface for indirect prompt injection.
Ingestion points: PDF data is ingested via scripts/extract_form_field_info.py, scripts/extract_form_structure.py, and libraries like pdfplumber and pypdf as described in SKILL.md and FORMS.md.
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore embedded instructions within the processed PDF content.
Capability inventory: The skill possesses capabilities for file writing via scripts/fill_fillable_fields.py and executing shell commands for PDF manipulation and image processing like qpdf, pdftotext, pdftk, and magick as outlined in SKILL.md and FORMS.md.
Sanitization: No explicit sanitization or validation of the extracted PDF content is performed before it is used to drive further agent actions.
[COMMAND_EXECUTION]: The skill relies on several command-line utilities for PDF processing, such as pdftotext, qpdf, pdftk, and magick. These tools are executed using arguments often derived from the analysis of external PDF documents.
[EXTERNAL_DOWNLOADS]: The OCR_AND_CONVERSION.md file specifies that the easyocr library downloads required language models upon initial execution. This is a standard and documented feature of the library used for its multi-language support.

pdf