Hi!
MCP Inspector already retrieves and displays an MCP server’s tool definitions, schemas, and connection behavior.
I’m building AgentSpec, a lightweight API that turns that tool surface into an advisory, machine-readable readiness summary before an agent connects to or a developer publishes a server.
It can surface:
- broad or risky tool permissions
- destructive or external side effects
- missing or unclear input schemas
- marketplace metadata gaps
Would an optional readiness-summary export or hook be useful in Inspector?
The goal would be advisory only: no default network call, no change to the existing inspection flow. A developer could explicitly request a structured report after Inspector has retrieved the server’s tool list.
I’m looking for honest feedback on whether this is useful during MCP server review and what output would be most valuable.
Demo: https://agentspec-one.vercel.app/
OpenAPI: https://agentspec-one.vercel.app/openapi.json
Hi!
MCP Inspector already retrieves and displays an MCP server’s tool definitions, schemas, and connection behavior.
I’m building AgentSpec, a lightweight API that turns that tool surface into an advisory, machine-readable readiness summary before an agent connects to or a developer publishes a server.
It can surface:
Would an optional readiness-summary export or hook be useful in Inspector?
The goal would be advisory only: no default network call, no change to the existing inspection flow. A developer could explicitly request a structured report after Inspector has retrieved the server’s tool list.
I’m looking for honest feedback on whether this is useful during MCP server review and what output would be most valuable.
Demo: https://agentspec-one.vercel.app/
OpenAPI: https://agentspec-one.vercel.app/openapi.json