local-seo
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The instruction to "follow [a separate file] as the active workflow" redirects the agent's logic to external directives.
- [PROMPT_INJECTION]: The skill processes untrusted business names and URLs, creating a surface for indirect prompt injection. Ingestion points: Business names and URLs provided via arguments. Boundary markers: None identified. Capability inventory: Diagnosis of GBP, NAP consistency, and local landing-page quality. Sanitization: No validation or sanitization of input data is present.
Audit Metadata