Loading ToolBench
Code topology tools for LLMs - an MCP server providing code navigation tools for LLMs. It parses source files and provides tools to index symbols, read/write definitions, and find references across codebases.
Four tools with basic descriptions and visible input schemas, but significant gaps in parameter documentation and output schema clarity. Tool names follow verb_noun convention (index, read_definition, write_definition, find_references), which is good. However, parameter descriptions are minimal, 'Directory path to index' and 'Symbol name to read' lack detail on format, constraints, or error cases. No output schemas documented. Error handling guidance absent. Descriptions range 60-90 chars, meeting minimum length but lacking LLM-optimization depth (target 50-200 chars with WHEN/WHY context). No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles. Missing pagination guidance for potentially large results (find_references, index).
Find all references to a symbol across the codebase
Index a directory and extract all code symbols (functions, types, classes, etc.) with their line ranges
Read the definition of a specific symbol from a file
Write or update a symbol definition in a file
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields without knowing response structure.
Parameter descriptions lack format/constraint details. 'Directory path to index' does not specify: absolute vs relative, symlink handling, max depth, or error cases (nonexistent path, permission denied).
No pagination or result-limiting guidance. 'index' and 'find_references' could return thousands of symbols; no mention of limits, pagination, or truncation behavior.
Missing tool annotations despite clear risk profiles. write_definition is destructive (WRITE risk); index and find_references are read-only. Annotations (destructiveHint, readOnlyHint) help agents reason about safety.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
No error recovery guidance. Descriptions do not explain what to do if symbol not found, file unreadable, or path invalid. LLMs receive no actionable next steps.