MCP Server for architectural analysis using OpenRouter and ASP.NET Core backend
Single tool 'check_my_architecture' has a Zod schema with all parameters optional and described, but critical issues undermine quality. Description is in Portuguese and lacks clarity on when/why to use the tool. Parameter descriptions are present but vague (e.g., 'Codigo opcional' instead of explaining what source code format is expected). No output schema is documented, the tool returns ArchitectureAnalysisResponse interface but this is not exposed to the LLM. Error handling returns generic error messages without recovery guidance. The tool combines multiple concerns (file reading, backend API call, optional auto-fix with file writes) in one operation, violating single-responsibility principle. No input validation or constraint documentation (e.g., file path traversal protection is implemented but not documented for the LLM). Tool name 'check_my_architecture' lacks a clear action verb and is ambiguous about what 'check' means.
Le um arquivo local ou usa source_code, envia para backend ASP.NET Core e retorna analise de arquitetura.
Tool description is in Portuguese ('Le um arquivo local...') and lacks clarity on when to use it. Descriptions must be in English and explain WHAT the tool does, WHEN to use it, and WHAT it returns. Current description does not answer these questions.
No output schema documented. The tool returns ArchitectureAnalysisResponse (analysis, violations, suggestions, patterns, confidence, refactoredCode, architectureDiagram, metadata) but LLMs cannot see this structure. Must document return fields and types in tool definition.
Tool combines multiple responsibilities: file reading, backend API call, optional auto-fix with file writes. This violates single-responsibility principle. Should split into separate tools: analyze_architecture (read-only) and apply_architecture_fix (write operation).
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 33 | <=2025-11-25 | v2 |
Error handling returns generic error messages without recovery guidance. E.g., 'Access denied: file is outside workspace' does not tell the LLM what to do next. Should include actionable recovery steps like 'Try with a file inside the workspace root' or 'Use source_code parameter instead of file_path'.
Tool name 'check_my_architecture' lacks a clear action verb. 'check' is vague, does it validate, analyze, scan, or audit? Should be 'analyze_architecture' or 'audit_architecture' to clearly signal intent to the LLM.