GA 自述的公开工作流示意

%% ==============================================================================
%% Google Antigravity (GA) 自述的公开工作流示意 (Public Workflow Concept)
%% 依据官方公开资料整理的概念流程，不代表未公开的模型内部推理。
%% 官方参考来源：
%% - https://antigravity.google/docs/agent/
%% - https://antigravity.google/docs/getting-started/
%% - https://antigravity.google/blog/introducing-google-antigravity/
%% ==============================================================================

flowchart TD
    %% 阶段一：用户配置与任务输入
    subgraph Stage1["1. 用户发起与上下文配置 (Project & Mode Selection)"]
        direction TB
        U1["用户提出任务需求<br/>(User Prompt / Task Input)"]
        U2["选择目标项目空间与工作模式<br/>(Active Project & Agent Mode)"]
        U1 --> U2
    end

    %% 阶段二：理解与计划
    subgraph Stage2["2. 上下文理解与计划生成 (Understanding & Planning)"]
        direction TB
        A1["理解任务目标与项目代码上下文<br/>(Context & Codebase Grounding)"]
        A2["生成结构化实施计划与任务清单<br/>(Plan & Task List Proposal)"]
        A1 --> A2
    end

    %% 阶段三：多工具协作执行
    subgraph Stage3["3. 多工具协同执行 (Tools Execution)"]
        direction TB
        T0["自主协作工具箱调度 (Autonomous Tool Usage)"]
        T1["编辑器 Editor<br/>(代码读取、跨文件修改、重构)"]
        T2["终端 Terminal<br/>(构建、依赖执行、运行命令)"]
        T3["浏览器 Browser<br/>(Web 交互、页面预览、控制台检查)"]
        T0 --> T1
        T0 --> T2
        T0 --> T3
    end

    %% 阶段四：产物生成
    subgraph Stage4["4. 过程与交付产物 (Artifacts Generation)"]
        direction TB
        AR0["生成高可读性结构化产物 (Artifacts)"]
        AR1["任务计划清单<br/>(Task Lists & Plans)"]
        AR2["代码变更比对<br/>(Code Diffs & Edits)"]
        AR3["视觉截图与界面排版<br/>(Screenshots & Previews)"]
        AR4["实施全流程演练<br/>(Walkthroughs & Docs)"]
        AR0 --> AR1
        AR0 --> AR2
        AR0 --> AR3
        AR0 --> AR4
    end

    %% 阶段五：自主验证
    subgraph Stage5["5. 结果验证与质量保障 (Verification)"]
        direction TB
        V1["按任务检查工作结果与功能<br/>(Task-appropriate Verification)"]
    end

    %% 阶段六：用户审查与回馈循环
    subgraph Stage6["6. 人机协作审查与迭代交付 (Review & Iteration)"]
        direction TB
        R1{"用户检查工作成果与验证产物<br/>(User Review & Evaluation)"}
        R2["Agent 结合用户反馈持续调整迭代<br/>(Iterative Refinement)"]
        R3["达成预期目标，正式交付与归档<br/>(Task Completed & Delivered)"]
    end

    %% 支线：Knowledge 可复用经验
    subgraph StageKnowledge["可复用知识库支线 (Knowledge System)"]
        direction TB
        K1[("项目知识与规范经验沉淀<br/>(Reusable Knowledge & Rules)")]
    end

    %% 主流程连线
    Stage1 --> Stage2
    Stage2 --> Stage3
    Stage3 --> Stage4
    Stage4 --> Stage5
    Stage5 --> R1

    %% 审查分支连线
    R1 -- "需要调整 / 补充反馈" --> R2
    R2 --> Stage2
    R1 -- "满意确认 / 验收通过" --> R3

    %% 知识沉淀与反哺虚线
    R3 -. "沉淀本次成功经验与规则" .-> K1
    R2 -. "记录调整经验与项目偏好" .-> K1
    K1 -. "为后续任务提供经验与规则上下文" .-> A1

    %% 节点样式
    classDef default font-family:sans-serif,font-size:13px;
    classDef userClass fill:#e8f0fe,stroke:#1a73e8,stroke-width:2px,color:#174ea6;
    classDef agentClass fill:#fef7e0,stroke:#f9ab00,stroke-width:2px,color:#b06000;
    classDef toolClass fill:#e6f4ea,stroke:#137333,stroke-width:2px,color:#0d652d;
    classDef artifactClass fill:#f3e8fd,stroke:#9334e6,stroke-width:2px,color:#681da8;
    classDef verifyClass fill:#fce8e6,stroke:#ea4335,stroke-width:2px,color:#c5221f;
    classDef knowledgeClass fill:#e0f2fe,stroke:#0284c7,stroke-width:2px,stroke-dasharray: 4 4,color:#0369a1;

    class U1,U2 userClass;
    class A1,A2,R2 agentClass;
    class T0,T1,T2,T3 toolClass;
    class AR0,AR1,AR2,AR3,AR4 artifactClass;
    class V1,R1,R3 verifyClass;
    class K1 knowledgeClass;