The Agent Skills Directory

[SAFE]: The skill demonstrates high security standards by implementing timingSafeEqual for bearer token validation to prevent timing attacks and emphasizing the use of secret environment variables for all sensitive credentials.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of a remote endpoint that processes untrusted tool call requests from AI clients.
Ingestion points: The netlify/functions/mcp.ts handler parses HTTP request bodies from external clients.
Boundary markers: Uses zod for strict input schema validation and provides specific guidance on writing tool descriptions to constrain model behavior.
Capability inventory: The server enables arbitrary tool execution based on external input, but the skill advises the principle of least privilege and provides patterns for audit logging.
Sanitization: Input validation is enforced via zod schemas for all tool parameters.
[EXTERNAL_DOWNLOADS]: The skill references standard, well-known libraries such as @modelcontextprotocol/sdk and zod via official package registries.
[COMMAND_EXECUTION]: Instructions include standard CLI commands for setting up development environments and connecting clients using netlify-cli and npx utilities.

netlify-mcp-servers