The Agent Skills Directory

[PROMPT_INJECTION]: The skill extracts text and metadata from user-provided PDF files, creating a surface for indirect prompt injection if the content contains malicious instructions.
Ingestion points: Text extraction occurs in SKILL.md (via pypdf and pdfplumber), scripts/extract_form_structure.py, and scripts/extract_form_field_info.py. OCR-based extraction using pytesseract is also supported.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified when presenting extracted text to the agent.
Capability inventory: The skill can write to the local filesystem (saving PDFs and images) and executes several CLI tools via subprocesses.
Sanitization: There is no evidence of filtering or sanitization of the extracted PDF content before it is processed by the agent.
[COMMAND_EXECUTION]: The skill relies on executing external command-line utilities for advanced PDF operations.
Evidence: SKILL.md, forms.md, and reference.md contain instructions for using pdftotext, qpdf, pdftk, pdftoppm, pdfimages, and magick (ImageMagick) to merge, split, and render PDF files. These tools are executed as subprocesses using the agent's environment capabilities.

pdf