seo-cannibalization-detector

Pass

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is entirely instructional and follows standard SEO best practices. It does not attempt to access sensitive files, execute shell commands, or perform unauthorized network operations.
  • [PROMPT_INJECTION]: The skill processes user-provided pages, which constitutes an indirect prompt injection surface. Ingestion points: Content from pages provided for cannibalization analysis in SKILL.md. Boundary markers: No explicit markers are defined. Capability inventory: No dangerous tools (shell access, file system writes, or non-whitelisted network calls) are available to the agent. Sanitization: No sanitization is performed. This surface is evaluated as safe due to the absence of exploitable system capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 20, 2026, 09:55 AM
Security Audit — agent-trust-hub — seo-cannibalization-detector