geo-content-optimizer

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of existing product descriptions, marketplace listings, and buyer questions to perform content rewrites. This creates a potential surface where adversarial instructions hidden in the source content could attempt to influence the agent's behavior.
  • Ingestion points: The skill collects "existing content, platform, verified product facts and policies" from external sources (SKILL.md).
  • Boundary markers: The workflow requires recording an "evidence boundary" and identifying vague or unsupported facts before rewriting (SKILL.md).
  • Capability inventory: The skill performs content transformation, structures data, and generates change logs.
  • Sanitization: No explicit sanitization or escaping of the input content is described, though the "Domain rules" mandate factual accuracy and adherence to evidence.
  • [EXTERNAL_DOWNLOADS]: The skill provides installation instructions using the npx skills command to fetch the ecommerce-seo-geo-skills package from the author's repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 05:39 AM
Security Audit — agent-trust-hub — geo-content-optimizer