mcp-server-patterns
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references the installation of the official
@modelcontextprotocol/sdkand thezodvalidation library via npm. These are standard, well-known dependencies for the intended purpose. - [COMMAND_EXECUTION]: The documentation describes how to implement tools that may execute commands or search functionality, which is a core feature of the Model Context Protocol. It explicitly recommends using Zod for strict input schema validation to ensure safety.
- [PROMPT_INJECTION]: The skill addresses the ingestion of external data by providing a framework for building MCP tools and resources. It implements security best practices by documenting the following: 1. Ingestion points: Data enters the system through tool input parameters defined in
inputSchemaand resource identifiers (URIs). 2. Boundary markers: The skill emphasizes defining strict input schemas for every tool to delimit expected data. 3. Capability inventory: Describes the registration of tools (registerTool) and resources (registerResource) across the SDK implementation. 4. Sanitization: Recommends the use of Zod for input validation and the return of structured error messages instead of raw stack traces to prevent data leakage.
Audit Metadata