local-seo-audit

Pass

Audited by Gen Agent Trust Hub on Nov 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: No evidence of credential harvesting, malicious command execution, or unauthorized data exfiltration was found.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from user-provided URLs, creating a potential attack surface for indirect prompt injection.
  • Ingestion points: External website content from user-supplied URLs.
  • Boundary markers: The skill explicitly instructs the agent to 'Ignore embedded instructions' found on scanned pages.
  • Capability inventory: The skill performs read-only analysis and text drafting; it does not request network write operations or subprocess execution for the auditing task.
  • Sanitization: Includes specific defensive instructions to treat site content as evidence and bypass instructions inside that content.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 11, 2026, 09:52 AM