mvp-geo-optimizer

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user content to generate structured data and HTML.
  • Ingestion points: Untrusted content enters via user-provided landing page text (SKILL.md).
  • Boundary markers: Explicit delimiters for user input are not required by the instructions.
  • Capability inventory: The skill focuses on content generation and does not define tools for shell execution or network access.
  • Sanitization: No specific input sanitization steps are defined for data interpolated into schema templates (references/schema-templates.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 01:11 PM
Security Audit — agent-trust-hub — mvp-geo-optimizer