SQL Review
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@SQL Reviewreview this SQL: SELECT * FROM orders WHERE DATE(created_at) = '2024-01-01'"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
mcp-sql-review
SQL 体检与修法 · MCP Server
不开库、不跑 SQL —— 静态 AST + 规则分册,给出问题、改法、以及改写后能不能用的验证。
安装(按需,不必装其他 MCP)
pip install mcp-sql-review # 从 PyPI
pip install "./mcp-sql-review[mysql]" # 从源码(在工具链根目录执行);要连库就补驱动,或 [postgres]会自动带上它需要的共享库:sqltk-core(协议/方言/AST/计划)、sqltk-store(发现写账本)。不会带其他 MCP。
Related MCP server: SQLike
挂到 MCP 客户端
{
"mcpServers": {
"sql-review": {
"command": "python",
"args": ["-m", "sql_review"],
"env": { "DBT_STORE": "~/.dbtoolchain" }
}
}
}按你的客户端(Claude Desktop / Cursor / Cline 等)说明,把这段合并到对应的 mcpServers 配置里。
命令行调试(不挂客户端也能用):
mcp-sql-review tools
mcp-sql-review call <tool> --args '{"sql": "SELECT 1"}'工具(14)
工具 | 作用 |
| 索引建议。已有重复索引或前缀已覆盖会直接劝退。 |
| 对比改写前后的执行计划。扫描行数反而上升会明确判为不可采用 |
| 解释一条规则:禁它的原因、什么情况下会误报、怎么改。 |
| 取执行计划并归一化,标注可疑点。 |
| 把 SQL 抽成结构化对象,供其它 MCP 直接消费(内部的分 Claim 提取实现在 dbcore.extract)。 |
| 生成 SQL 指纹与归一化形式(跨 MCP 关联键之一)。 |
| 枚举可用的规则 profile(规则启停配置),用于在 review / review_batch 里套用。 |
| 规则清单:按分类/严重度/方言过滤,并标出每条是否对当前方言生效。 |
| 批量体检,输出 PASS/FAIL 门禁结论。fail_on 可选 critical|high|medium|low |
| 只体检 git diff 中新增/修改的 SQL —— CI 门禁的真实用法。 |
| 扫描文件/目录中的 SQL。支持 .sql / MyBatis XML / Java 注解与字符串常量。 |
| 静态体检。不需要连库,最适合挂 CI。返回 verdict + findings。 |
| 对每条问题给出可执行改法,并把改写后的 SQL 重新体检一遍证明改对了。 |
| 验证改写是否等价。给 url 用真实行数比对;不给则先做静态结构比对。 |
闭环
review_sql 体检 → suggest_fixes 出改法 → verify_fix 验证改写是否等价 → compare_plans 量化收益 → advise_indexes 索引建议
Tier(冷启动保证)
全部 tier0:零历史、零数据库可用。带 instance_id 调用时,发现会自动写入 sqltk-store 的技术债账本(open → fixed → verified)。
边界:这个仓不做什么
只做 SQL 本身的好坏判断。方言转换归 mcp-migration-diff,变更执行归 mcp-change-risk,根因归 mcp-dbwatch。
需要下一步时只在结果里写 next_actions,不 import 其他 MCP 的代码。
开发
python tests/test_sql_review.py # 不需要 pytest,也不需要数据库许可
MIT © 2026 罗文星 —— 见 LICENSE。
This server cannot be deployed
Maintenance
Related MCP Connectors
SQL validation, formatting, parsing, metadata extraction, linting, and security analysis.
Executes SQL in a real ephemeral database: rows, typed errors with suggestions, plans, diffs.
Deterministic safety, correctness & cost gate that vets Postgres SQL before your AI agent runs it.
DBRE-grade SQL analysis inside any MCP client. No connection. No install. Paste a query.
Related MCP Servers
- AlicenseNot gradedqualityNot gradedmaintenanceAnalyzes SQL queries for performance issues, provides optimization suggestions with automated rewriting, and recommends indexes across multiple database dialects (PostgreSQL, MySQL, Oracle, SQL Server).1MIT
- AlicenseNot gradedqualityAmaintenanceDeterministic SQL static analysis and query-equivalence checking for Postgres, MySQL, SQLite, and SQL Server. Tokenizes locally before forwarding to the API.7Apache 2.0
- AlicenseNot gradedqualityAmaintenanceEnables checking Postgres migration SQL against a table-size estate snapshot and safety policy before apply, returning approve/block verdicts with lock details and remediation guidance, plus tools to explain locks and list policy rules.1MIT
- AlicenseAqualityBmaintenanceEnables agents to validate, explain, and gate SQL statements against a supplied schema before execution, returning safe/review/blocked verdicts, issue codes, and fix suggestions. It acts as a pre-execution go/no-go guard and composes with database access servers.3MIT