nutrient-document-processing

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation describes an alternative setup using npx to download and execute the official @nutrient-sdk/dws-mcp-server package from npm. This is a standard deployment method for the vendor's MCP integration.
  • [PROMPT_INJECTION]: The skill's document processing capabilities (OCR and text extraction) create an indirect prompt injection surface where instructions embedded in a processed document could attempt to influence the agent's behavior.
  • Ingestion points: User-supplied documents in various formats (PDF, DOCX, etc.) processed via the build API.
  • Boundary markers: No explicit instruction delimiters or boundary markers are suggested for handling the extracted text.
  • Capability inventory: The skill utilizes network access to api.nutrient.io and local file system access for document management.
  • Sanitization: The documentation focuses on functionality; users should ensure they validate or sanitize extracted text if it is used for downstream decision-making.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 10:24 PM
Security Audit — agent-trust-hub — nutrient-document-processing