AI-native full-stack framework with compiler-driven code generation for building AI applications with agents, memory, MCP integration, and UI components
Critical gaps across all dimensions. Tools are inferred from example-app.py rather than explicitly registered with schemas. No input schemas visible for any tool except 'query' (partial). Descriptions are present but generic (10-50 chars, well below 50-200 char baseline for LLM optimization). No parameter descriptions. No output schemas documented. No error handling guidance. Tool names lack clarity (e.g., 'search' vs 'search_code'). This is a framework demo, not a production MCP server implementation.
Get repository information via GitHub MCP server
Execute SQL queries against PostgreSQL database via postgres MCP server
Read file contents via filesystem MCP server
Search the web via web_search MCP server
Search code repositories via GitHub MCP server
No input schemas visible for any tool. Tools are inferred from MCPTool() calls in example-app.py, not explicitly registered with JSON Schema. Only 'query' shows a partial schema (sql parameter).
Descriptions are generic and under 50 characters (well below 50-200 char baseline). E.g., 'Search code repositories via GitHub MCP server' (47 chars) lacks WHEN to use, prerequisites, or return value guidance.
No parameter descriptions visible. 'query' parameter in search_code and search tools has a description, but get_repository, query (SQL), and read_file parameters lack any guidance. LLMs cannot infer parameter meaning from names alone.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 36 | 2026-07-28+ | v2 |
No output schemas documented. Tools return unstructured results (e.g., 'repository information', 'SQL query results'). LLMs cannot plan downstream tool calls or extract required fields without knowing the response structure.
Tool 'search' is ambiguous, does not start with a clear verb and conflicts with 'search_code'. Naming convention should be verb_noun (e.g., 'search_web'). LLMs conflate similar names and waste reasoning cycles.