mcp-integration
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and examples follow security best practices for developing AI agent extensions. No malicious patterns were identified.
- [EXTERNAL_DOWNLOADS]: The documentation provides setup examples referencing official packages from the @modelcontextprotocol organization, which are trusted resources.
- [INDIRECT_PROMPT_INJECTION]: The skill identifies the surface for data ingestion from external servers and correctly recommends mitigations such as strict tool allow-lists and input validation.
- Ingestion points: External MCP servers (SSE, HTTP, WebSocket).
- Boundary markers: Pre-allowed tool lists in command frontmatter are recommended.
- Capability inventory: External service connectivity (databases, APIs, file systems).
- Sanitization: Guidance provided for input validation and error handling.
- [DYNAMIC_EXECUTION]: Documentation describes the headersHelper functionality for dynamic authentication, which is a standard feature for supporting advanced authentication flows.
Audit Metadata