MCP Server using Spring Boot Java for stock price management and user data operations
Server has 9 tools with minimal schema documentation and weak descriptions. Tool names follow verb_noun convention (good), but descriptions are generic (10-30 chars, well below 50-200 char baseline for A+ tools). Input schemas visible but lack detail: parameters have types but most descriptions are trivial ('The user ID', 'The search query'). No output schemas documented. No error handling guidance. No parameter constraints (enums, ranges, formats). Critical gaps in composition: getAllUsers and getAllUsersDefault are redundant; updateUser accepts Map<String, Object> with no validation hints. No evidence of idempotency, pagination metadata, or chaining IDs in responses. Security: no audit logging visible, no permission gates on destructive tools (deleteUser). Overall: definitions are skeletal, suitable for internal use but not production-grade agent tooling.
Add a new user
Delete a user
Get all users
Get all users with default pagination
Get stock quote for a symbol
Get a single user by ID
List available stock symbols
Search for users by query
Descriptions are generic and under 20 characters for most tools (e.g., 'Get all users', 'Get a single user by ID'). LLMs cannot distinguish when to use getAllUsers vs getAllUsersDefault vs getUserById without richer context.
updateUser accepts Map<String, Object> with no schema, validation hints, or examples. LLMs cannot infer which fields are valid, required, or mutable.
getAllUsers and getAllUsersDefault are redundant. Per composition pattern, avoid multiple tools doing the same thing differently. This forces LLM to reason about which to call, wasting tokens.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 44 | 2026-07-28+ | v2 |
Update a user
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields.
deleteUser has no confirmation step, dry-run, or permission gate. Destructive operations should support confirmation to prevent accidental data loss. No audit logging visible.