pdf-to-markdown

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data (PDF files) which could contain embedded instructions intended to influence agent behavior.
  • Ingestion points: The script scripts/pdf_to_markdown.py reads content from user-specified PDF files using the pymupdf4llm library.
  • Boundary markers: The extracted text is output directly to the agent without specific delimiters or markers to distinguish untrusted content from system instructions.
  • Capability inventory: The skill possesses the ability to read and write files on the local filesystem as directed by the user.
  • Sanitization: No explicit sanitization or filtering of extracted text is performed within the scripts.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the uv package manager to automatically retrieve necessary dependencies at runtime.
  • Evidence: Usage of uv run --with pymupdf4llm and uv run --with pymupdf in SKILL.md ensures required libraries are available.
  • Context: These references target well-known libraries on the official Python Package Index (PyPI) for PDF processing tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 02:29 PM
Security Audit — agent-trust-hub — pdf-to-markdown