nanoorabot – AI DBA Danny 开源发布

项目地址：

https://github.com/valenwong-exa/nanoorabot

Enjoy by yourself ！

基于nanobot和nanobot-webui，MIT lisence！

各位警界同僚，请注册Github，并打个星。另外欢迎大家一起vibe coding，在github提交创意，一起创造.

安装请参考INSTALL.md, DBA请参考DBA_BEST_PRACTICE.md

一些用法：

自动补丁

知识库

自愈

连接不上监听，自己解决

DBA的日常任务

该项目争取周更，主要在做这几件事

1. 测试和上传更多的Oracle 和 Linux Skill。

2. 测试和上传更多的MCP Server

3. 整合Oracle 26ai的向量检索标准化MCP Server

4. 更新原版nanobot对系统管理有益的更新到目前的版本

5. 测试一个自主运行DBA的SOP工作清单

6. 完善审计和危险命令防御

7. 本地读取pdf和office文件（更新中），当大模型不支持时，切换为本地。

8. tokens统计（更新中）

9. 多DBA协作界面（计划中）

10. 一些本身bug的处理

11. 语音输入指令（只做了界面）

12. 图像绘制MCP添加

该项目在原有项目的基础上添加了

1. 主机Dashboard

2. 历史命令审计

3. 忘却记忆

4. 提交图片到多模态模型进行分析

5. win2linux openssh工具

6. 灵魂和外貌

7. sqlplus工具

8. 绑定了oracle mcp，测试了rag mcp

Danny – DBA 1号

付鹏 – 系统专家

瑞秋 – 数据分析师

张捷 – 架构师

Nanoorabot – AI System Agent

本项目是一个基于 nanobot 与 nanobot-webui 二次开发的 AI System Agent 机器人系统，目标是构建一个能够长期运行、具备自主执行能力的智能系统管理员。This project is an AI System Agent robot system developed based on nanobot and nanobot-webui, aiming to build an intelligent system administrator capable of long-term operation and autonomous execution.

*本项目由资深DBA系统架构师创建和维护，依赖vibe coding 创建。欢迎大家一起vide coding。 *This project is created and maintained by a senior DBA system architect, with development powered by vibe coding. Everyone is welcome to join in and do vibe coding together.

它面向主机管理、数据库管理和自动化运维场景，强调可扩展、可观测、可持续运行。It targets host management, database management, and automated operations and maintenance (O&M) scenarios, emphasizing scalability, observability, and sustainable operation.

特别是针对 Oracle 数据库的运维，本系统提供了深度的自动化管理、故障排查与诊断能力。Especially for Oracle database O&M, this system provides deep automated management, troubleshooting, and diagnostic capabilities.

系统特别引入了Oracle 26ai 的向量库能力作为知识库，大幅增强了专业知识的检索准确度。The system specifically introduces the vector database capabilities of Oracle 26ai as its knowledge base MCP Server, significantly enhancing the retrieval accuracy professional knowledge.

系统采用Oracle MCP Server作为Oracle数据库的访问工具The system specifically introduces Oracle MCP Server to access Oracle database.

通过整合 AI、工具 (Tool)、记忆 (Memory) 与 Web UI 的架构，我们形成了一个可落地的 Agent 管理平台，但本项目在开发中，仅仅限于用于DEV或者UAT环境，如果用于生产环境，后果自负，本项目不承担任何责任。。By integrating the architecture of AI, Tools, Memory, and Web UI, we have formed a practical Agent management platform. This project is currently under development and is only intended for use in the DEV/UAT environment. If it is used in the production environment, the user shall bear all the consequences.This project assumes no responsibility. *目前本项目仅仅测试了KIMI和Deepseek API *Currently we just tested with KIMI and Deepseek API

核心特性 / Core Features

自主运行的系统管理员 / Autonomous System Administrator

: 能够理解自然语言指令并将其转化为实际的运维操作。Capable of understanding natural language commands and translating them into actual O&M operations.
Oracle 数据库深度运维 / Deep Oracle Database O&M

: 内置针对 Oracle 数据库的专属管理技能，实现高效的日常维护与异常处理。Built-in exclusive management skills for Oracle databases to achieve efficient daily maintenance and exception handling.
26ai 向量知识库 / 26ai Vector Knowledge Base

: 依托 26ai 向量库构建强大的专业领域知识库，赋予 AI 准确的行业经验参考。Relying on the 26ai vector database to build a powerful professional domain knowledge base, endowing AI with accurate industry experience references.
双轨记忆机制 / Dual-track Memory Mechanism

: 支持灵活的短期会话记忆与长期的历史经验积累，支持通过 /forget-session 等显式指令进行灵活的记忆管理。Supports flexible short-term session memory and long-term historical experience accumulation, and supports flexible memory management through explicit commands such as /forget-session.

基于项目 / Based On

本项目的基础能力得益于以下优秀的开源项目：The foundational capabilities of this project benefit from the following excellent open-source projects:

nanobot
nanobot-webui

内置第一个AI系统管理员 Danny | Buildin the first AI system administrator Danny

主要的系统管理员特性 | Key Features for SYSTEM Admin

删除记忆 | Memory Deletion Commands Summary

因为运维机器人可能不需要记忆 Because OPS robot may not needs memory

可管理主机和数据库列表 | Server Host and Database list

AI交付记录审计存储在26ai数据库

数据库和主机危险命令防御

内置Oracle SkiLL

来自AI Data Flatform 团队整理

感谢整理和提供

https://github.com/oracle/skills

https://github.com/krisrice/oracle-db-skills

skill-name work-type

ora-alert-log-analysis daily
ora-awr-reports daily
ora-connection-pooling daily
ora-dbms-scheduler daily
ora-health-monitor daily
ora-index-strategy daily
ora-locking-concurrency daily
ora-optimizer-stats daily
ora-plsql-dynamic-sql daily
ora-redo-log-management daily
ora-rman-backup-recovery daily
ora-space-management daily
ora-sql-tuning daily
ora-top-sql-queries daily
ora-undo-management daily
ora-wait-events daily
ora-adrci-usage demand
ora-advanced-queuing demand
ora-ash-analysis demand
ora-auditing demand
ora-database-links demand
ora-database-testing demand
ora-dataguard demand
ora-data-masking demand
ora-data-modeling demand
ora-dotnet-oracle demand
ora-dynamic-sql demand
ora-edition-based-redefinition demand
ora-encryption demand
ora-explain-plan demand
ora-inmemory-column-store demand
ora-java-oracle-jdbc demand
ora-json-in-oracle demand
ora-materialized-views demand
ora-memory-tuning demand
ora-metadata-analysis demand
ora-migrate-db2-to-oracle demand
ora-migrate-mongodb-to-oracle demand
ora-migrate-mysql-to-oracle demand
ora-migrate-postgres-to-oracle demand
ora-migrate-redshift-to-oracle demand
ora-migrate-snowflake-to-oracle demand
ora-migrate-sqlite-to-oracle demand
ora-migrate-sqlserver-to-oracle demand
ora-migrate-sybase-to-oracle demand
ora-migrate-teradata-to-oracle demand
ora-migration-assessment demand
ora-migration-cutover-strategy demand
ora-migration-data-validation demand
ora-multitenant demand
ora-network-security demand
ora-nodejs-oracledb demand
ora-online-operations demand
ora-oracle-apex demand
ora-oracle-cloud-oci demand
ora-oracle-migration-tools demand
ora-oracle-text demand
ora-ords-architecture demand
ora-ords-authentication demand
ora-ords-auto-rest demand
ora-ords-file-upload-download demand
ora-ords-installation demand
ora-ords-metadata-catalog demand
ora-ords-monitoring demand
ora-ords-pl-sql-gateway demand
ora-ords-rest-api-design demand
ora-ords-security demand
ora-partitioning-strategy demand
ora-pl-sql-best-practices demand
ora-plsql-code-quality demand
ora-plsql-collections demand
ora-plsql-compiler-options demand
ora-plsql-cursors demand
ora-plsql-debugging demand
ora-plsql-error-handling demand
ora-plsql-package-design demand
ora-plsql-patterns demand
ora-plsql-performance demand
ora-plsql-security demand
ora-plsql-unit-testing demand
ora-privilege-management demand
ora-python-oracledb demand
ora-rac-concepts demand
ora-rman-basics demand
ora-row-level-security demand
ora-schema-migrations demand
ora-sequences-identity demand
ora-spatial-data demand
ora-sql-best-practices demand
ora-sqlcl-basics demand
ora-sqlcl-cicd demand
ora-sqlcl-data-loading demand
ora-sqlcl-ddl-generation demand
ora-sqlcl-formatting demand
ora-sqlcl-liquibase demand
ora-sqlcl-mcp-server demand
ora-sqlcl-scripting demand
ora-sql-injection-avoidance demand
ora-sql-patterns demand
ora-sql-property-graph demand
ora-tablespace-design demand
ora-transaction-management demand
ora-user-management demand
ora-version-control-sql demand
ora-virtual-columns demand
ora-xml-in-oracle demand
ora-erd-design other
ora-exadata-features other
ora-golang-oracle other

DBA routine work SOP

基于Oracle 26ai的RAG文档维护和向量搜索功能

连接Oracle MCP Server

许可证 / License

本项目采用 MIT 许可证开源。This project is open-sourced under the MIT License.

*请保留Valen Wang（王探长）,Wanbin, Lijie和AI Database Plaftform Team的署名 Please retain the bylines of Valen Wang and the AI Database Platform Team.

任何人都可以免费获取本软件及其相关文档的副本，并在满足 MIT 协议条款的前提下进行使用、复制、修改、合并、发布、分发等操作。Anyone can obtain a copy of this software and its associated documentation files for free, and use, copy, modify, merge, publish, distribute, etc., under the terms of the MIT License.