document-extraction-api
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes documents provided by the user, which could contain instructions intended to influence the agent's behavior. Ingestion points: Local files passed to scripts/parse.py via the --input argument. Boundary markers: The script returns structured JSON or Markdown text, which clearly separates document content from agent instructions. Capability inventory: The skill can read local files, perform network calls to the vendor's API, and write output files. Sanitization: Extracted text is provided without specific filtering for injection patterns.
- [EXTERNAL_DOWNLOADS]: The skill requires the nutrient-dws Python package, which is the official SDK from the vendor (PSPDFKit).
- [COMMAND_EXECUTION]: The skill instructions guide the agent to execute its own parsing script (scripts/parse.py) and includes cost-estimation logic to ensure user confirmation for expensive document processing tasks.
Audit Metadata