canonical-tag

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides standard SEO implementation patterns for HTML, Next.js, and server configurations (Apache, Nginx). No suspicious network operations, obfuscation, or remote code execution detected.
  • [INDIRECT_PROMPT_INJECTION]: The skill reads local project context files (.claude/project-context.md, .cursor/project-context.md) to extract site configuration.
  • Ingestion points: Reads project-specific context files for URL and language structure.
  • Boundary markers: None specific, but the usage is limited to identifying project metadata.
  • Capability inventory: No subprocess calls, network writes, or file-writing capabilities are present in the skill instructions.
  • Sanitization: Standard instructional interpretation by the AI agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 12:26 PM
Security Audit — agent-trust-hub — canonical-tag