tokendiet-mcp provides 4 tools with clear, action-oriented names (read, run, search, fetch) and reasonable descriptions (100-150 chars each). All tools have input schemas with typed parameters and enums where appropriate. However, output schemas are not documented in the visible code, parameter descriptions lack depth (e.g., 'mode' enum values not explained), and error handling guidance is absent. The 'run' tool poses security risks (IRREVERSIBLE, shell execution) with no dry-run or confirmation pattern. Tool composition is sound (each does one thing), but descriptions could better explain WHEN to use each tool vs. alternatives.

• Document output schemas for all tools. Specify field names, types, and whether results are paginated. Example: 'Returns {content: string, tokens: number, mode_used: string}'.
• Add dry-run support to 'run' tool or wrap it in a confirmation pattern. E.g., add 'dry_run: boolean' param; if true, return what would execute without executing.
• Expand parameter descriptions with actionable guidance. For 'mode' enum, explain: 'auto: intelligently choose based on file type; full: return entire content uncompressed; outline: return code structure only; symbol: extract specific function/class'.
• Add error recovery guidance to tool descriptions. E.g., 'If file not found, try search() to locate it. If search returns no results, check glob pattern syntax.'
• Clarify tool selection in descriptions. E.g., 'read: for local files; fetch: for remote URLs. search: to find files by pattern before reading.'