This MCP server provides a set of tools for code analysis, enabling fast and precise symbol lookup. When using these tools, you should first call start_analyzer to initialize the analyzer, passing the root directory of the code as a parameter. Once the analyzer is started, other interfaces can be invoked. The find_definition tool provides functionality to locate the position where a variable or function is defined. It is typically used during code analysis when you need to inspect the implementation of a called function. This tool requires a variable name or function name as a parameter. The find_references tool enables locating the positions where a variable or function is referenced or used. It is commonly used to determine where a specific function is called during code analysis. Like find_definition, this tool also requires a variable name or function name as a parameter.
Critical gaps across all dimensions. Tool names lack action verbs (start_analyzer, stop_analyzer are weak; find_definition/find_references are acceptable). Descriptions are present but minimal (10-40 chars, below the 50-200 char LLM-optimized baseline). Parameter descriptions exist but are generic ('function name or variable name'). No output schemas documented, tools return list[str] with no field structure. No error handling guidance. No idempotency guarantees. No pagination for potentially large result sets. Tools appear to be inferred from Pydantic models rather than explicit MCP tool registration, capping per-tool scores at 50 maximum.
Find definition position of a symbol
Find all references of a symbol
Start the code analyzer in a workspace
Stop the code analyzer
No output schemas documented. Tools return list[str] with no field structure, field names, or guidance on what each string represents. LLMs cannot parse results or chain to downstream tools.
Parameter descriptions are generic and lack format/constraint guidance. 'function name or variable name' does not explain expected format, case sensitivity, or how to handle symbols in nested scopes.
Tool descriptions are too short (10-40 chars, below 50-char minimum). 'Start the code analyzer in a workspace' lacks context on when to call it, what it initializes, or prerequisites.
No error handling or recovery guidance. Tools return bare list[str] on success but no documented error cases, timeout behavior, or what to do if symbol not found.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 36 | <=2025-11-25 | v2 |
Tool definitions inferred from Pydantic models, not explicit MCP tool registration visible in source. Cannot verify actual tool registration, capability negotiation, or protocol compliance.