Three tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. All tools are read-only (low risk), but descriptions lack actionable context for LLM selection. Parameter descriptions are minimal (avg ~30 chars vs baseline 72 chars). No output schemas documented. Missing error recovery guidance. Naming is verb-forward and clear, but parameter constraints are underspecified.

• Document output schemas for all three tools. Example for search_files: 'Returns {results: [{path: string, name: string, type: 'file'|'directory', matchType: 'name'|'content'}], totalCount: number, limit: number}'. This enables LLMs to chain results into list_directory or analyze_code.
• Expand parameter descriptions with format and constraint details. E.g. 'path: Directory path relative to codebase root (e.g. src/components). Must exist and be readable. Defaults to root if omitted.' This prevents invalid inputs.
• Add WHEN-to-use guidance to tool descriptions. E.g. search_files: 'Use when looking for specific files by name or content. For browsing directory structure, use list_directory instead. Searches recursively by default.'
• Document enum option meanings. For analyze_code type param: 'imports: list module/package imports; functions: extract function signatures; classes: extract class definitions; all: return all three categories.'
• Add error recovery hints. E.g. 'If no results found, try a broader query or use list_directory to explore the structure. If analyze_code fails on a file, it may be binary, use list_directory to verify file type first.'