上下文管理：Rules + RAG + Codebase + MCP 四要素
05 / Solution detail
AI Coding Assistant Solution (CodeBuddy)
A multi-end AI coding assistant built with Tencent CodeBuddy, spanning IDE/Plugin/CLI. Includes code completion, project-aware Q&A, the Craft development agent, AICR code review, and Repo Wiki - all anchored in a three-ring AICoding methodology.
Solution contents
00
Overview
Deliverables
AI Coding Assistant Solution (CodeBuddy)
1CodeBuddy multi-end product (IDE/Plugin/CLI)
2AICoding three-ring methodology
3Craft task-level development agent
4AICR code review and security agents
5Repo Wiki and custom RAG
5
Deliverables
6
delivery chapters
01
Multi-end IDE / Plugin / CLI Architecture
Multi-end access under unified management, with weekly actives of 7,000+ IDE, 50,000+ Plugin, and 2,000+ CLI users.
Capability points
1AI IDE：编码-编译-调试全链路上下文感知，原生支持Spec+TDD
2插件：支持 VS Code、JetBrains 全系、Visual Studio、Xcode、独家支持微信开发者工具
3CLI：贴合 Linux 开发理念，前端轻负载，自动化与并发任务友好
4多端接入 × 统一管理
5当前周活：IDE 7,000+、Plugin 5W+、CLI 2,000+
02
AICoding Methodology - Three-ring Closed Loop
任务流管理：TaskPlan 任务拆解 + Spec 规格文档 + TDD 测试驱动
事务管理：Craft Agent + Repo Wiki Agent + AICR Agent + 安全 Agent
来自腾讯内部大规模落地的 SOTA 实践
AI 输出贴合项目背景、方案明确、可直接执行
A closed loop of context, task flow, and transaction - AI outputs that fit the project background, with clear plans and ready-to-execute steps.
03
Task-level Craft Development Agent
Smart requirement analysis, planning, and AI-driven coding with self-testing - specialized in toolchains, events, SDK integration, and shader authoring.
- 智能分析需求并制定开发计划
- 由 AI 完成编码与测试验证
- 擅长工具链开发、事件开发、SDK 集成、Shader 编写
- 两种使用模式：独立完成任务 / 协助打草稿
- 上下文清晰且可快速验证的工具或原型场景
04
AICR Code Review & Security Agents
AICR runs both local and repo-level reviews; the security agent integrates Tencent TCA, covering defects, standards, vulnerabilities, and dead code.
- AICR：基于自定义规则的 MR 目的性代码审查，支持 GIT 协议与 API 接入
- 本地代码评审 + 代码仓库级评审双场景
- 安全智能体集成腾讯 TCA，覆盖代码缺陷/规范/漏洞/无效代码
- 沙盒环境后端自动修复 + IDE 人工选择修复双模式
- 已交付英雄联盟、腾讯新闻等 10+ 大型团队
05
Repo Wiki Architecture Agent & Custom RAG
Real-time Git monitoring auto-generates code docs; custom RAG supports multiple data sources with @knowledge-base retrieval augmentation.
01
Repo Wiki：实时监听 Git 仓库变更，自动生成与持续更新代码知识文档
Core
02
自动生成架构图、业务流程图、组件图、依赖关系图、类图与数据流图
03
自定义 RAG：代码语法语义分析技术，可结合业务逻辑切片与存储
04
多数据源支持：.md / .docx / .pdf 与 .zip / .tar 等压缩包
Implementation flow
06
Enterprise Service System & Infrastructure
Capability points
01
服务体系：产品 + 实施 + 咨询 全链路
02
AI 编码培训：腾讯内部最佳实践 + 场景化技巧
03
定制开发：MCP / 研发工具集 / 企业报表
04
知识体系建设：企业私有知识库（SDK / 业务 / 编码规范）
05
基础设施：国产化算力 / NVIDIA 算力 / 腾讯云 GPU 支持
End-to-end product + implementation + consulting, with AI coding training, custom MCP/tooling/reporting, and compute from domestic, NVIDIA, and Tencent Cloud.
5
Capability points
06
Implementation analysis
Contact
Book an enterprise AI rollout review
Tell us about your scenario, data conditions, and current AI usage. We will help you identify the right rollout path.