The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides mechanisms to read and extract text from untrusted external PDF files, creating a surface for indirect prompt injection attacks.
Ingestion points: PDF files processed via pdfplumber, pypdf, and pytesseract in SKILL.md snippets.
Boundary markers: Absent. The instructions do not define delimiters or provide warnings to the agent to disregard instructions found within extracted PDF text.
Capability inventory: The skill utilizes file system write access (wb mode) for saving manipulated PDFs and displays extracted content to the agent.
Sanitization: Absent. Extracted data is processed as raw text without validation or filtering.
[COMMAND_EXECUTION]: The OCR functionality uses pytesseract and pdf2image, which are wrappers that invoke external command-line tools (Tesseract OCR and Poppler) installed on the host system.
[EXTERNAL_DOWNLOADS]: The skill suggests the installation of several third-party libraries from standard registries, including pypdf, pdfplumber, reportlab, pytesseract, pdf2image, weasyprint, camelot, pdfrw, EasyOCR, and playwright.

pdf-processing