Single tool 'web_search' with significant quality gaps. Description is in Chinese (non-standard for LLM tooling) and lacks English clarity. Input schema present but parameters lack proper type constraints and validation guidance. No output schema documented. No error handling guidance. Tool naming is acceptable but descriptions do not meet LLM-optimization baselines (194 chars avg for A+ tools). Parameter descriptions are minimal and lack format/constraint details required for reliable LLM invocation.

• Translate tool description to English and expand to 100-200 chars explaining WHAT (web search), WHEN (need real-time info), and WHAT IT RETURNS (search results with sources). Current description is 89 chars in Chinese.
• Document output schema: define fields returned (e.g., results[], each with title, url, snippet, source, timestamp). Specify result limit (e.g., max 10 results) and pagination if applicable.
• Add format constraints to 'query' parameter: min 1 char, max 200 chars, no special operators. Add guidance to 'explanation' parameter: must start with verb (获取/了解/查询/确认) and be 10-100 chars.
• Add error handling guidance: 'If search fails due to network timeout, retry after 5 seconds. If rate limited (>100 queries/min), wait 60 seconds. If query is empty, return error: "Query required, min 1 char."'
• Add tool annotations to schema: include 'readOnlyHint: true' and 'idempotentHint: true' to signal to LLM that tool is safe to retry and does not modify state.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry