The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): The skill processes untrusted external data (PDF files) and possesses capabilities that can be abused if malicious instructions are found within those files.
Ingestion points: Uses PdfReader, pdfplumber.open, and convert_from_path to read local PDF documents.
Boundary markers: No delimiters or 'ignore' instructions are present to prevent the agent from obeying instructions embedded in extracted text.
Capability inventory: The skill allows writing files (writer.write, to_excel, c.save, doc.build) and executing shell commands via various PDF utilities.
Sanitization: Extracted content is not sanitized or validated before being printed or processed.
[Command Execution] (MEDIUM): The skill relies on and provides instructions for several command-line utilities (pdftotext, qpdf, pdftk, pdfimages), which necessitates the agent executing shell commands with parameters often derived from input file names.

pdf