Two tools with complete input schemas and descriptions, but significant gaps in output documentation and parameter clarity. boss_search and boss_filters both have descriptions (194 and 156 chars respectively, within baseline 34-392 range) and typed parameters. However, output schemas are only documented in prose within descriptions, not as formal JSON Schema. Parameter descriptions lack constraint details (e.g., city code format, page bounds, filter_params structure). No error handling guidance. Tool names follow verb_noun pattern (search, filters) but descriptions embed example values ('Python', '101210100') which LLMs may reuse literally. Missing idempotency and retry guidance for read-only operations.

• Add formal JSON Schema definitions for return types. Define job object schema with fields: name (string), salary (string), company (string), city (string), labels (array), jobId (string), securityId (string). Document sorts, nlpFilters, labelFilter structures in boss_filters return.
• Replace example values with enum constraints. For sort parameter, define enum: [-1, 1, 2] with descriptions. For city, document format as 'China weather-station code (9 digits, e.g. 101210100 for Hangzhou)' without embedding the example.
• Add bounds to numeric parameters: page (1-100), sort (enum or range), and document filter_params as 'object with keys from boss_filters response (e.g. {"salary": "407", "experience": "104,105"})'.
• Add error handling guidance: 'Returns error if city code is invalid, session expired (t2 token), or network unreachable. If auth fails, regenerate session via boss login command.'
• Document idempotency: 'Both tools are read-only and safe to retry. Repeated calls with same parameters return identical results.'