Loading ToolBench

Reading the current corpus…

spring-ai-otel-example MCP server · F · ToolBench

spring-ai-otel-example

github.com/kycasdzxc/spring-ai-otel-example

A Spring Boot application demonstrating Spring AI MCP client integration with OpenTelemetry observability, featuring a chat interface that uses MCP tools for Grafana integration

<=2025-11-25STREAMABLE_HTTP1 tools

13/100

Definition Quality6

Protocol Readiness32

SupportabilityNot measured-

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: Logging

Deprecated protocol patterns detected

Summary

This is not an MCP server, it is a Spring Boot application that CONSUMES MCP servers as a client. The single 'tool' listed ('MCP Tools (Dynamic)') is not a tool definition but a description of how the app dynamically loads tools from external MCP servers via Spring AI's ToolCallbackProvider. No tool schemas, descriptions, or parameter definitions are visible in the source code. The app exposes a /chat endpoint and /tools endpoint that delegate to external MCP servers; it does not define or register any tools itself. Tool definitions are inferred from external MCP server responses, not explicitly declared in this codebase.

Tools (1)

MCP Tools (Dynamic)read onlyauth7/100

MCP tools provided by external MCP servers, dynamically loaded and exposed through Spring AI's ToolCallbackProvider

Issues (5)

CRITICALtool ↗

Not an MCP server, this is an MCP client application. Tool definitions are dynamically loaded from external MCP servers, not defined in this codebase. No explicit tool schemas, parameter descriptions, or input validation are visible.

MCP Tools (Dynamic)

CRITICALtool ↗

Tool definitions are inferred from external MCP server responses. No explicit tool registration with schemas is visible in ChatController.java or any other source file.

MCP Tools (Dynamic)

CRITICALconstrained-input ↗

No input parameter schemas, type definitions, or constraints are declared for any tool in the codebase.

MCP Tools (Dynamic)

HIGHrecovery-guide ↗

No error handling guidance or recovery instructions visible. The /chat endpoint returns generic HTTP errors without actionable context for the LLM.

MCP Tools (Dynamic)

Recommendations

Clarify that this is an MCP client application, not an MCP server. If the goal is to expose MCP tools via HTTP, implement an MCP server that wraps the Spring AI client and registers tools with explicit schemas.
If wrapping external MCP tools, add a schema validation and enrichment layer that documents each tool's parameters, constraints, and error recovery paths in the HTTP response.
Implement structured error responses that guide the LLM on recovery: 'Tool X failed: <reason>. Try Y instead.' instead of raw HTTP status codes.
Add per-tool metadata (required permissions, idempotency, destructiveness) to the /tools endpoint response so clients can make informed decisions about tool selection.
Document the MCP server connection configuration (host, port, transport) and add health checks to verify external MCP servers are reachable before exposing their tools.

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Logging (deprecated) - log to stderr or use OpenTelemetry

Score history

Overall score trend

First recorded score · v2 rubric

13/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	13	<=2025-11-25	v2