WeMCP Agent - An agent system with support for MCP servers, skills, and agents with a FastAPI backend and Next.js frontend
WeMCP has a single tool 'multiply' with minimal definition quality. The tool has a basic description (18 characters: 'Multiplies two numbers together.'), which falls short of the 10-1024 character guideline and lacks operational context. Input schema is present with typed parameters (a: float, b: float) and parameter descriptions, but the output schema is not documented. The tool name follows verb_noun convention (multiply is clear), but lacks context about when/why to use it. No error handling guidance is provided. The tool appears in a test file rather than being part of a formal server definition, raising questions about whether this is the actual production tool or merely an example. The server implements only 1 tool total, which is extremely limited for a general-purpose agent platform. Prompts and resources are declared but tool definition depth is poor.
Multiplies two numbers together.
Tool description is too short (18 chars). Guideline requires 10-1024 chars with context for when/why to use the tool. Current description lacks operational guidance, prerequisites, and return value documentation.
Output schema is not documented. LLMs cannot infer what fields the tool returns or plan downstream operations. Critical for agent composition and error recovery.
No error handling or recovery guidance. If the tool fails (e.g., numeric overflow, invalid input), there is no message telling the LLM what to do next or how to self-correct.
Tool appears only in a test file (backend/tests/fastmcp_server_test.py), not in explicit production server registration. This suggests the tool definition is inferred or example-based rather than formally declared.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 47 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 44 | - | v1 |
No parameter constraints or validation rules documented. Input description for 'a' and 'b' is simply 'First number' and 'Second number', no guidance on valid ranges, precision, or edge cases (e.g., overflow, negative numbers, zero handling).
Single tool server. A production agent platform with only one tool (multiply) is severely limited. No indication of planned tools for real-world agent workflows (CRUD, search, communication, integration).