market-segments

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external data sources (market studies, customer databases, research documents) without using boundary markers or delimitation.
  • Ingestion points: User-provided market studies and customer databases (SKILL.md).
  • Boundary markers: The instructions lack delimiters or specific 'ignore instructions' warnings for external content.
  • Capability inventory: The skill's primary focus is analytical output; no dangerous command execution or network tools are specifically invoked in the instructions.
  • Sanitization: No content sanitization or validation logic is specified for the input data.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 02:37 PM
Security Audit — agent-trust-hub — market-segments