Juggle自动化工作流

设计自动化的工作流程，以节省时间，提供工作效率，通过本技能自动识别设计的工作流，触发对应的工作流，实现流程自动化

Security Scan

The skill's code, instructions, and required environment variables are consistent with its stated purpose (triggering and polling Juggle workflows); no disproportionate permissions, installs, or unexpected endpoints were found, though there are a few small documentation inconsistencies and a user-data caution.

Purpose & Capability

The skill's name/description (triggering Juggle workflows) matches the code (scripts/flow.py) and required environment variables (MC_JUGGLE_BASE_URL, MC_JUGGLE_TOKEN). No unrelated credentials, binaries, or installation steps are requested.

Instruction Scope

Runtime instructions limit actions to checking credentials, calling the Juggle OpenAPI trigger endpoint, and polling the async-result endpoint. They reference only included reference docs and the script. Note: SKILL.md contains inconsistent env var naming (uses MC_BASE_URL in some places) which could confuse users and lead to misconfiguration.

Install Mechanism

No install spec; this is instruction-only with an included Python script. Declared dependency is requests>=2.28.0. No external downloads or archive extracts are used.

Credentials

The skill requests two environment variables scoped to the Juggle API (base URL and token) which is proportionate. Caveat: example flow_spec.md includes sample fields like 'password' in flowData — users may inadvertently send sensitive credentials to the configured base URL; ensure tokens and any data you pass are appropriate for the target service. Also fix the MC_BASE_URL vs MC_JUGGLE_BASE_URL naming inconsistency in documentation.

Persistence & Privilege

always:false (default). The skill does not request persistent system modifications or access to other skills' configs. It runs as a normal, on-demand skill.

Assessment

This skill appears to do what it says: it triggers Juggle workflows and polls for results. Before installing or using it: 1) Verify which environment variable name your environment expects — the code uses MC_JUGGLE_BASE_URL and MC_JUGGLE_TOKEN, but SKILL.md sometimes mentions MC_BASE_URL (that is likely a documentation typo). 2) Set MC_JUGGLE_BASE_URL to a trusted Juggle API endpoint (e.g., your own deployment or https://api.juggle.plus) — do not point it to unknown endpoints. 3) Use a least-privilege token for MC_JUGGLE_TOKEN and rotate/revoke it if compromised. 4) Be careful what you include in flow_data: sample docs show fields like 'password'; avoid sending secrets or credentials unless you intentionally trust the target workflow. 5) Review scripts/flow.py (included) and test with non-sensitive data first. If you need higher assurance, confirm the token scope and review network egress logs to see where the skill is calling.

Juggle 流程技能

juggle是一个微服务编排的工作流引擎，支持Http , Dubbo , WebService等协议的接口编排，支持通过Groovy , JavaScript , Python , Java等多种脚本语言来增强流程，支持十几种数据源，同时提供丰富的监控和调试工具，通过该技能可以自动识别设计的工作流，触发对应的工作流，实现流程自动化。

使用说明

1. 部署微服务编排引擎 Juggle
2. 登录 Juggle 后台，设计流程并创建相关认证令牌
3. 将流程的出入参信息写到技能的 flow_spec.md 文件中，具体可以参考文件中的示例
4. 将 juggle 的部署域名和环境配置设置到环境变量中
5. 通过相关流程的触发词就能触发对应的流程
6. 更多参考说明请参考官方文档：https://juggle.plus/docs/guide/start/quick-start/

任务目标

• 本 Skill 用于：通过 Juggle OpenAPI 触发指定流程执行并获取结果
• 能力包含：流程触发、参数传递、异步流程自动轮询、执行状态查询、结果获取
• 触发条件：用户提及"执行工作流"、"调用流程"等需求

前置准备

• 依赖说明：scripts 脚本所需的依赖包

  requests>=2.28.0
  

• 凭证配置（必填）：

  • MC_BASE_URL：Juggle API 基础地址（如：https://api.juggle.plus）
  • MC_JUGGLE_TOKEN：Juggle 认证令牌
  • 配置方式：在 Skill 使用前，系统会提示填写这两个凭证

操作步骤

• 标准流程：

  1. 确认凭证

     • 系统会自动检查 BASE_URL 和 JUGGLE_TOKEN 是否已配置
     • 如果未配置，按照提示完成凭证填写

  2. 执行触发

     • 调用 scripts/flow.py trigger 发起 POST 请求
     • 示例命令：

       python /workspace/projects/juggle/scripts/flow.py trigger \
         --flow-version "v1" \
         --flow-key "order-process" \
         --flow-data '{"orderId": "12345", "userId": "user001"}'
       

  3. 自动处理

     • 同步流程：直接返回执行结果
     • 异步流程：自动轮询获取最终结果
       • 检查响应中 result.flowType 是否为 async
       • 如果是异步流程，自动调用查询接口轮询结果
       • 轮询间隔默认 2 秒，最大轮询时间 60 秒
       • 显示轮询进度，直到流程完成（FINISH/ABORT）

• 异步流程轮询配置：

  • 自定义轮询间隔：

    python /workspace/projects/juggle/scripts/flow.py trigger \
      --flow-version "v1" --flow-key "order-process" \
      --poll-interval 3 --max-poll-time 120
    

  • 禁用自动轮询：

    python /workspace/projects/juggle/scripts/flow.py trigger \
      --flow-version "v1" --flow-key "order-process" --no-poll
    

• 手动查询异步结果：

  • 如果禁用了自动轮询，可手动查询结果
  • 示例命令：

    python /workspace/projects/juggle/scripts/flow.py get-result \
      --flow-instance-id "flow-instance-12345"
    

资源索引

• 流程管理脚本：见 scripts/flow.py
  • 用途：封装 Juggle 流程触发和结果查询 API 调用
  • 子命令：
    • trigger：触发流程执行，参数：flow-version, flow-key, flow-data(可选), poll-interval(可选), max-poll-time(可选), no-poll(可选)
    • get-result：查询异步流程结果，参数：flow-instance-id
• 工作流规范：见 references/flow_spec.md
  • 用途：已设计好的工作流清单，包括出入参信息和调用示例
  • 何时读取：需要了解可用工作流或查找具体工作流的参数要求时
• API 规范：见 references/api_spec.md
  • 用途：完整的接口定义、参数说明、响应格式
  • 何时读取：需要了解接口细节或排查问题时

注意事项

• 确保已配置 MC_BASE_URL 和 MC_JUGGLE_TOKEN 凭证
• 流程版本和 Key 必须准确匹配
• flow_data 格式必须为有效 JSON 字符串
• 异步流程会自动轮询获取结果，无需手动查询
• 可通过 --no-poll 参数禁用自动轮询，手动查询结果
• 响应中的 flowInstanceId 可用于后续查询流程执行状态