Server has 3 tools with minimal schema documentation. All tools lack parameter type definitions in visible code, only string types are inferred from descriptions. Tool descriptions are present but generic (50-80 chars), below the 194-char baseline for production tools. No output schemas documented. No error handling guidance. No tool annotations (readOnlyHint, destructiveHint). Parameter descriptions are minimal. Tools appear to be search-only (READ_ONLY risk), but descriptions do not explicitly state this or guide recovery on failure. Spring AI MCP framework is used, but tool registration via MethodToolCallbackProvider suggests inferred definitions rather than explicit schema registration, capping per-tool scores at 50 maximum.

• Add explicit JSON Schema definitions with 'type', 'minLength', 'maxLength', and 'pattern' for searchText parameter. E.g., searchText: {type: 'string', minLength: 1, maxLength: 256, description: 'Search keyword (1-256 chars). Supports partial matches.'}
• Expand tool descriptions to 150 - 250 chars. Include: (1) what the tool does, (2) when to use it vs. alternatives, (3) what it returns, (4) error recovery. E.g., 'Search interview questions from 面试鸭 database. Use this for technical interview prep. Returns up to 20 matching questions with difficulty and tags. If no results, try broader keywords or check searchArticles for tutorials.'
• Document output schema for each tool. Specify fields (e.g., questionSearch returns [{id, title, difficulty, tags, url}]). Include total count and pagination info (page, pageSize, hasMore).
• Add pagination parameters: limit (1 - 100, default 20) and offset (default 0). Return total count and next_cursor in response. Cap results at 20 - 50 items per call to prevent context bloat.
• Add tool annotations: mark all three tools with readOnlyHint=true (they do not modify state). This signals to agents that these calls are safe to retry and do not have side effects.