Critical failures across all dimensions. Descriptions are minimal (10-26 chars), below the 20-char floor for scoring above 0-20. No parameter documentation exists. Tool names lack clarity (e.g., 'foreground_required' is ambiguous, does it check, set, or request foreground?). No output schemas documented. No error handling guidance. The codebase references Java/Android implementation files, not TypeScript MCP server code, suggesting tool definitions are inferred rather than directly visible in source. Average of 8 tools with these failures yields ~12.
Get device information
Indicate that foreground execution is required
Get current time
Fetch content from web
Search the web
List workspace contents
Read workspace file
Write to workspace file
All 8 tools have empty input schemas ({}). LLMs cannot infer parameter types, constraints, or required fields.
All tool descriptions are 10 - 26 characters, below the 20-char floor. Descriptions lack WHAT the tool does, WHEN to use it, and what it returns.
Tool 'foreground_required' has ambiguous naming. Does it check if foreground is required, set a flag, or request foreground execution? Name does not start with a clear action verb (get_, set_, check_, request_).
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 33 | 2026-07-28+ | v2 |
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields (e.g., does workspace_list return file paths, sizes, timestamps?).
Tool definitions appear inferred from Java/Android implementation files, not directly visible in TypeScript MCP server source. No explicit MCP tool registration found in provided code.