VJSP CodeVJSP Code is an innovative full-stack intelligent programming assistant that deeply integrates AI with visual development to reshape an efficient programming experience. Installation Launch VS Code Quick Open ( Ctrl+P), paste the following command, and press enter.Copied to clipboard |
VJSP Code 全栈智能编程助手VJSP Code — Full-Stack Intelligent Programming Assistant产品介绍 | Product IntroductionVJSP Code 是一款面向研发全流程的全栈智能编程助手，以自然语言交互为核心，融合 AI 编程、智能交互、任务管理、内置工具集、扩展与自定义、自动化及多模态处理能力，帮助开发者快速生成、理解、优化并交付高质量代码。VJSP Code 覆盖代码生成、代码补全、快速编辑、代码注释、单元测试、文档生成等开发全流程，内置多种智能工作模式和自定义模式，支持开放生态，覆盖软件开发过程，AI 驱动开发，提升编程效率。 VJSP Code is a full-stack intelligent programming assistant for the entire development lifecycle. Centered on natural-language interaction, it integrates AI programming, intelligent interaction, task management, a built-in toolset, extension & customization, automation, and multimodal processing — helping developers rapidly generate, comprehend, optimize, and deliver high-quality code. As a standalone IDE ready out of the box, VJSP Code covers the full development flow of code generation, completion, quick editing, comments, unit testing, and documentation, with multiple built-in and custom intelligent modes, an open ecosystem, and AI-driven development that boosts programming efficiency. 主要特点 | Key Highlights
人工智能编程 | AI Programming代码生成 Code Generation通过自然语言描述需求，自动生成符合项目规范的函数、模块、类或代码片段。 Generate functions, modules, classes, or code snippets from natural-language descriptions, in line with project conventions. 代码补全 Code Completion提供行级、函数级的单行或多行代码推荐，支持通过注释生成代码。 Provide line- and function-level single-/multi-line suggestions, and generate code from comments. 快速编辑 Quick Edit选中代码即可通过自然语言完成代码修复与代码重构。 Fix and refactor selected code quickly through natural language. 代码注释 Code Comments为函数或逐行代码自动生成规范注释，支持自定义注释模板与格式。 Auto-generate standard comments for functions or individual lines, with customizable templates and formats. 代码操作 Code Operations以自然语言驱动代码的定位、修改与应用，生成代码一键应用并形成差异对比。 Drive code locate, modify, and apply operations via natural language, applying generated code in one click with diff comparison. 智能问答 AI Q&A面向研发领域优化的技术问答与知识查询，精准解释项目代码，帮助快速熟悉项目。 Development-optimized technical Q&A and knowledge lookup; explain project code precisely for quick onboarding. 智能模式 Intelligent Modes内置代码、问答、计划、调试、协调者五种智能工作模式，支持手动切换与自定义模式编排。 Five built-in modes — Code, Q&A, Plan, Debug, Coordinator — with manual switching and custom mode orchestration. 任务规划 Task Planning自动将复杂需求拆解为多级子任务清单（To-Do List），支持调整顺序与分步执行。 Automatically decompose complex requirements into a multi-level To-Do List, supporting reordering and step-by-step execution. 建议回复 Suggested Replies任务完成后自动生成多条可一键执行的下一步操作建议。 After a task completes, generate multiple one-click suggestions for the next steps. 自动批准 Auto-Approval按权限规则预配置的工具无需人工确认即可执行，同时保留高危操作拦截机制。 Pre-authorized tools run without manual confirmation, while high-risk operations remain intercepted. 检查点 Checkpoints在关键节点自动创建快照，支持一键回退至任意历史检查点。 Automatically create snapshots at key milestones, with one-click rollback to any checkpoint. 提示词 Prompt Management支持增强提示与提示词策略管理，自动补全上下文约束、输出格式与项目规范。 Support prompt enhancement and prompt-strategy management, auto-completing context constraints, output format, and project conventions. 能力扩展与开放 | Capability Extension & OpennessSkills 技能 Skills基于开放 AgentSkills 标准，通过 SKILL.md 描述可复用的领域任务技能，支持全局/项目级与模式专属技能。 Reusable domain-task skills based on the open AgentSkills standard via SKILL.md, supporting global/project-level and mode-specific skills. MCP 生态 MCP兼容 Model-Context-Protocol 标准，对接数据库、GitLab 等外部服务，支持多服务与全局/项目两级配置。 Compatible with the Model-Context-Protocol standard to connect external services such as databases and GitLab, with multi-server and global/project-level configuration. 智能工作流 Intelligent Workflows结合任务规划、子 Agent 调度、Skills 与 MCP 实现复杂工程端到端流程，支持条件分支、循环执行与 AI 动态调整。 Combine task planning, sub-agent dispatch, Skills, and MCP for end-to-end flows, with branching, loops, and AI-driven step adjustment. 工具调用 Tool Calling标准化 Agent 工具调度内核，统一管理内置工具、Skills 与 MCP 工具，负责参数校验、权限拦截与结果回传。 A standardized agent tool-dispatch kernel that unifies built-in, Skills, and MCP tools, handling validation, permission interception, and result feedback. 上下文 Context基于打开文件、项目结构等上下文提供精准建议，支持 @ 引用文件、函数、类、变量。 Provide accurate suggestions based on open files, project structure, and other context; support @ references to files, functions, classes, and variables. 记忆库 Memory Base通过 AGENTS.md 持久化项目上下文，使 AI 跨会话理解技术栈、历史决策与任务状态。 Persist project context via AGENTS.md, so the AI understands the tech stack, past decisions, and task state across sessions. 代码库索引 Codebase Index对代码库建立索引，结合向量检索实现全仓库语义搜索，快速理解项目结构与代码关系。 Index the codebase and use vector search for repository-wide semantic search, quickly comprehending project structure and code relationships. 大型项目 Large Projects借助代码库索引与全局上下文感知，在大型复杂工程中高效导航、检索与理解代码。 With codebase indexing and global context awareness, efficiently navigate, search, and understand large, complex projects. 规则 / 流程 Rules & Processes支持自定义规则（项目级/全局级、外部 .md 文件、glob 加载）、指令、工作流与温度调节，可约束敏感文件访问与编码标准。 Support custom rules (project/global levels, external .md files, glob loading), instructions, workflows, and temperature control; enforce sensitive-file access and coding standards. 系统集成 System Integration通过 MCP 与工具系统对接数据库、GitLab 等企业服务，支持本地模型部署与多模型接入。 Connect enterprise services such as databases and GitLab via MCP and the tool system, with local-model deployment and multi-model access. 多模态处理 Multimodal Processing支持视觉设计稿识别、报错截图 OCR 分析与文档图片结构化处理。 Support visual-draft recognition, error-screenshot OCR analysis, and document-image structuring. 本地模型与多模型 Local & Multi-Model支持接入 OpenAI 兼容接口的本地模型，实现内网离线运行；会话内可随时切换模型。 Connect local models with OpenAI-compatible APIs for offline intranet operation; switch models anytime within a session. 自动化启动 Automated Startup打开工作区时自动执行环境检测、代码库初始化引导与预设模式/工作流加载。 Automatically run environment checks, repository initialization guidance, and preset mode/workflow loading when a workspace opens. |