Support for SSE, STDIO in MySQL MCP server, includes CRUD, database anomaly analysis capabilities. Supports multiple transport modes and includes database health analysis, index optimization, and SQL execution capabilities.
This server has significant definition quality gaps. Of 5 tools, 3 have descriptions under 100 characters (execute_sql, get_chinese_initials, get_table_index), and all lack proper English descriptions (Chinese descriptions do not count for LLM-optimization criteria). Tool naming is mostly acceptable (verb_noun convention), but descriptions are too terse to guide LLM selection. All tools declare input schemas with types, but parameter descriptions are single-line summaries. Output schemas are not documented anywhere in the codebase. Error handling is not evident from the tool implementations provided. The execute_sql tool is especially risky, it accepts arbitrary SQL with no visible validation, dry-run support, or confirmation gate, yet the description is only 25 characters in Chinese.
在MySQL数据库上执行SQL (支持多条SQL语句,以分号分隔)
创建表结构时,将中文字段名转换为拼音首字母字段
获取当前连接的mysql库的索引使用情况,包含冗余索引情况、性能较差的索引情况、未使用索引且查询时间大于30秒top5情况(Get the index usage of the currently connected mysql database, including redundant index situations, poorly performing index situations, and the top 5 unused index situations with query times greater than 30 seconds)
获取当前mysql的健康状态(Analyze MySQL health status )
根据表名搜索数据库中对应的表索引,支持多表查询(Search for table indexes in the database based on table names, supporting multi-table queries)
All tool descriptions are primarily in Chinese with minimal English; LLM optimization requires 50-200 char English descriptions explaining WHAT, WHEN, and WHAT-TO-EXPECT. Current descriptions are 15-35 chars.
execute_sql tool is WRITE (destructive) with no visible dry-run, confirmation gate, or permission check. Agents can accidentally drop tables or corrupt data. No error handling guidance visible.
No output schemas documented for any tool. LLMs cannot plan downstream tool chains or extract required fields without knowing what each tool returns.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 49 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 29 | - | v1 |
Parameter descriptions are 7-20 characters in Chinese only, providing minimal guidance. E.g., 'query' param in execute_sql has no mention of multi-statement support, encoding, max length, or format constraints.
No visible input validation, error recovery guidance, or actionable error messages in tool implementations. E.g., 'table not found' or 'invalid SQL syntax' errors lack suggestions for recovery.
get_table_index and get_db_health_running descriptions do not distinguish when to use one vs. the other, both claiming database introspection. Overlapping responsibilities invite LLM confusion.