pdf-analysis

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run to invoke a local Python script (caption.py) for OCR tasks on image-based PDF pages. The script is called with a specific internal path and arguments are passed as a list, preventing shell injection vulnerabilities.
  • [PROMPT_INJECTION]: The skill processes untrusted PDF data which could contain malicious instructions designed to influence the agent's behavior (indirect prompt injection).
  • Ingestion points: Multiple functions (detect_pdf_type, extract_text_pdf, extract_tables_pdf, extract_scanned_pdf) read content from external PDF files.
  • Boundary markers: No delimiters or instructions are used to separate extracted document text from agent instructions.
  • Capability inventory: The skill can execute local scripts via subprocess and write temporary files to /tmp/.
  • Sanitization: No filtering or sanitization of the extracted PDF text is performed before it is passed to the agent context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 12:25 PM
Security Audit — agent-trust-hub — pdf-analysis