multimodal-analysis

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [NO_CODE]: The skill consists entirely of documentation and instructional text in markdown format. It does not include any scripts, executables, or package manifests that could perform system operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to interpret content from external, untrusted sources (PDFs, images, technical diagrams), which creates an attack surface for indirect prompt injection.
  • Ingestion points: Processing and extraction of data from external media files as described in SKILL.md.
  • Boundary markers: The instructions lack delimiters or explicit warnings for the agent to ignore instructions embedded within the processed media.
  • Capability inventory: The skill does not provide its own tools or scripts, relying instead on the agent's native multimodal capabilities.
  • Sanitization: There are no defined procedures for sanitizing, validating, or filtering the content extracted from the media files.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 08:05 AM
Security Audit — agent-trust-hub — multimodal-analysis