workflow-orchestrator

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Workflow Orchestrator

工作流编排器

A skill that automatically coordinates workflows across multiple skills, triggering follow-up actions at appropriate milestones.

一款可自动协调多Skill工作流，在合适的里程碑节点触发后续操作的Skill。

When This Skill Activates

该Skill的激活场景

This skill should be triggered automatically when:

A skill completes its main workflow
A milestone is reached (PRD complete, implementation done, etc.)
User says "complete workflow" or "finish the process"

当以下情况发生时，该Skill会自动触发：

某个Skill完成其主工作流
达到里程碑节点（如PRD完成、功能实现完成等）
用户说出“完成工作流”或“结束流程”

How It Works

工作原理

┌─────────────────────────────────────────────────────────────┐
│                    Workflow Orchestration                   │
├─────────────────────────────────────────────────────────────┤
│                                                             │
│  1. Detect Milestone → 2. Read Hooks → 3. Execute Chain    │
│                                                             │
│  prd-planner complete                                       │
│       ↓                                                     │
│  workflow-orchestrator                                      │
│       ↓                                                     │
│  ┌─────────────────────────────────────┐                   │
│  │ auto-trigger self-improving-agent   │ (background)       │
│  │ auto-trigger session-logger         │ (auto)            │
│  └─────────────────────────────────────┘                   │
│                                                             │
└─────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────┐
│                    Workflow Orchestration                   │
├─────────────────────────────────────────────────────────────┤
│                                                             │
│  1. Detect Milestone → 2. Read Hooks → 3. Execute Chain    │
│                                                             │
│  prd-planner complete                                       │
│       ↓                                                     │
│  workflow-orchestrator                                      │
│       ↓                                                     │
│  ┌─────────────────────────────────────┐                   │
│  │ auto-trigger self-improving-agent   │ (background)       │
│  │ auto-trigger session-logger         │ (auto)            │
│  └─────────────────────────────────────┘                   │
│                                                             │
└─────────────────────────────────────────────────────────────┘

Trigger Configuration

触发配置

Read trigger definitions from

skills/auto-trigger/SKILL.md

yaml

hooks:
  after_complete:
    - trigger: self-improving-agent
      mode: background
    - trigger: session-logger
      mode: auto
  on_error:
    - trigger: self-improving-agent
      mode: background

从

skills/auto-trigger/SKILL.md

读取触发定义：

yaml

hooks:
  after_complete:
    - trigger: self-improving-agent
      mode: background
    - trigger: session-logger
      mode: auto
  on_error:
    - trigger: self-improving-agent
      mode: background

Execution Modes

执行模式

Mode	Behavior	Use When
`auto`	Execute immediately, no confirmation	Logging, status updates
`background`	Execute without blocking	Reflection, analysis
`ask_first`	Ask user before executing	PRs, deployments, major changes

模式	行为	使用场景
`auto`	立即执行，无需确认	日志记录、状态更新
`background`	后台执行，不阻塞当前流程	反思、分析
`ask_first`	执行前询问用户	提交PR、部署、重大变更

Milestone Detection

里程碑检测

PRD Complete

PRD完成

markdown

Detected when:
- docs/{scope}-prd.md exists
- All phases in {scope}-prd-task-plan.md are checked
- Status shows "COMPLETE"

Actions:
1. Trigger self-improving-agent (background)
2. Trigger session-logger (auto)

markdown

检测条件：
- docs/{scope}-prd.md 文件存在
- {scope}-prd-task-plan.md 中的所有阶段均已勾选
- 状态显示为"COMPLETE"

执行操作：
1. 触发self-improving-agent（background模式）
2. 触发session-logger（auto模式）

Implementation Complete

功能实现完成

markdown

Detected when:
- All PRD requirements implemented
- Tests pass
- Code committed

Actions:
1. Trigger code-reviewer (ask_first)
2. Trigger create-pr if changes staged
3. Trigger session-logger (auto)

markdown

检测条件：
- PRD中的所有需求均已实现
- 测试通过
- 代码已提交

执行操作：
1. 触发code-reviewer（ask_first模式）
2. 若有暂存的变更，触发create-pr
3. 触发session-logger（auto模式）

Self-Improvement Complete

自我优化完成

markdown

Detected when:
- Reflection complete
- Patterns abstracted
- Skill files modified

Actions:
1. Trigger create-pr (ask_first)
2. Trigger session-logger (auto)

markdown

检测条件：
- 反思流程完成
- 已提炼模式与经验
- Skill文件已修改

执行操作：
1. 触发create-pr（ask_first模式）
2. 触发session-logger（auto模式）

Universal Learning (Any Skill Complete)

通用学习（任意Skill完成时）

markdown

Detected when:
- ANY skill completes its workflow
- User provides feedback
- Error or issue encountered

Actions:
1. Trigger self-improving-agent (background)
2. Trigger session-logger (auto)

The self-improving-agent:
- Extracts experience from completed skill
- Identifies patterns and insights
- Updates related skills with learned patterns
- Consolidates memory for future reference

markdown

检测条件：
- 任意Skill完成其工作流
- 用户提供反馈
- 遇到错误或问题

执行操作：
1. 触发self-improving-agent（background模式）
2. 触发session-logger（auto模式）

self-improving-agent执行内容：
- 从已完成的Skill中提取经验
- 识别模式与洞察
- 用学到的模式更新相关Skill
- 整合经验以供后续参考

Error Handling (on_error)

错误处理（on_error）

Detected when:

A command returns non-zero exit code
Tests fail after following skill guidance
User reports the guidance produced incorrect results

Actions:

Trigger self-improving-agent (background) for self-correction
Trigger session-logger (auto) to capture error context

当以下情况发生时触发：

命令返回非零退出码
遵循Skill指导后测试仍失败
用户反馈指导结果不正确

执行操作：

触发self-improving-agent（background模式）进行自我修正
触发session-logger（auto模式）捕获错误上下文

Hook Implementation in Skills

Skill中的钩子实现

To enable auto-trigger, add this section to any skill's SKILL.md:

markdown

undefined

要启用自动触发功能，需在任意Skill的SKILL.md中添加以下内容：

markdown

undefined

Auto-Trigger (After Completion)

自动触发（完成后）

When this skill completes, automatically trigger:

yaml

hooks:
  after_complete:
    - trigger: skill-name
      mode: auto|background|ask_first
      context: "relevant context"
  on_error:
    - trigger: self-improving-agent
      mode: background

当该Skill完成时，自动触发：

yaml

hooks:
  after_complete:
    - trigger: skill-name
      mode: auto|background|ask_first
      context: "相关上下文"
  on_error:
    - trigger: self-improving-agent
      mode: background

undefined

Current Skill Hooks

当前支持的Skill钩子

prd-planner: After PRD complete → self-improving-agent + session-logger
self-improving-agent: After improvement → create-pr + session-logger
prd-implementation-precheck: After implementation → self-improving-agent + session-logger
code-reviewer: After review → self-improving-agent + session-logger
debugger: After debugging → self-improving-agent + session-logger
create-pr: After PR created → session-logger
session-logger: No trigger (terminates chain)

prd-planner: PRD完成后 → self-improving-agent + session-logger
self-improving-agent: 优化完成后 → create-pr + session-logger
prd-implementation-precheck: 实现完成后 → self-improving-agent + session-logger
code-reviewer: 评审完成后 → self-improving-agent + session-logger
debugger: 调试完成后 → self-improving-agent + session-logger
create-pr: PR创建完成后 → session-logger
session-logger: 无后续触发（工作流链终止）

Universal Learning Pattern

通用学习模式

┌─────────────────────────────────────────────────────────────┐
│                  ANY Skill Completes                        │
└──────────────┬──────────────────────────────────────────────┘
               │
               ↓
    ┌──────────────────────┐
    │ workflow-orchestrator │
    └──────────┬───────────┘
               │
    ┌──────────┴─────────┐
    ↓                   ↓
self-improving-agent  session-logger
    ↓                   ↓
Learn from experience  Save context
    ↓                   ↓
Update skills         Log session
    ↓
create-pr (if modified)

undefined

┌─────────────────────────────────────────────────────────────┐
│                  ANY Skill Completes                        │
└──────────────┬──────────────────────────────────────────────┘
               │
               ↓
    ┌──────────────────────┐
    │ workflow-orchestrator │
    └──────────┬───────────┘
               │
    ┌──────────┴─────────┐
    ↓                   ↓
self-improving-agent  session-logger
    ↓                   ↓
Learn from experience  Save context
    ↓                   ↓
Update skills         Log session
    ↓
create-pr (if modified)

undefined

Workflow Examples

工作流示例

Example 1: PRD Creation Workflow

示例1：PRD创建工作流

User: "Create a PRD for user authentication"
        ↓
prd-planner executes
        ↓
Phase 6 complete: PRD delivered
        ↓
workflow-orchestrator detects milestone
        ↓
┌─────────────────────────────────┐
│ Background: self-improving-agent │ → Learns from PRD patterns
│ Auto: session-logger             │ → Saves session
└─────────────────────────────────┘

用户："为用户认证功能创建PRD"
        ↓
prd-planner 执行
        ↓
第6阶段完成：PRD交付
        ↓
workflow-orchestrator 检测到里程碑
        ↓
┌─────────────────────────────────┐
│ 后台执行：self-improving-agent │ → 从PRD模式中学习
│ 自动执行：session-logger             │ → 保存会话
└─────────────────────────────────┘

Example 2: Full Feature Workflow

示例2：全功能开发工作流

User: "Create a PRD and implement it"
        ↓
prd-planner → workflow-orchestrator
        ↓
self-improving-agent → workflow-orchestrator
        ↓
prd-implementation-precheck
        ↓
implementation complete → workflow-orchestrator
        ↓
code-reviewer → self-improving-agent → workflow-orchestrator
        ↓
create-pr → workflow-orchestrator
        ↓
session-logger

Each step triggers

self-improving-agent

to learn from the experience.

用户："创建PRD并实现该功能"
        ↓
prd-planner → workflow-orchestrator
        ↓
self-improving-agent → workflow-orchestrator
        ↓
prd-implementation-precheck
        ↓
功能实现完成 → workflow-orchestrator
        ↓
code-reviewer → self-improving-agent → workflow-orchestrator
        ↓
create-pr → workflow-orchestrator
        ↓
session-logger

每个步骤都会触发

self-improving-agent

从经验中学习。

Implementation Steps

实现步骤

Step 1: Detect Milestone

步骤1：检测里程碑节点

Check for completion indicators:

bash

undefined

检查完成标识：

bash

undefined

PRD complete?

PRD是否完成？

grep -q "COMPLETE" docs/{scope}-prd-task-plan.md

All phases checked?

所有阶段是否已勾选？

grep -q "^- [x].*Phase 6" docs/{scope}-prd-task-plan.md

PRD file exists?

PRD文件是否存在？

ls docs/{scope}-prd.md

undefined

ls docs/{scope}-prd.md

undefined

Step 2: Read Trigger Config

步骤2：读取触发配置

bash

undefined

bash

undefined

Read hooks from auto-trigger skill

从auto-trigger skill中读取钩子配置

cat skills/auto-trigger/SKILL.md

undefined

cat skills/auto-trigger/SKILL.md

undefined

Step 3: Execute Hooks

步骤3：执行钩子

For each hook in order (before_start, after_complete, on_error):

Check if condition is met
Execute based on mode
Pass context to triggered skill
Wait/continue based on mode

按顺序处理每个钩子（before_start、after_complete、on_error）：

检查条件是否满足
根据模式执行
将上下文传递给被触发的Skill
根据模式决定等待或继续

Step 4: Update Status

步骤4：更新状态

Log what was triggered and the result:

markdown

undefined

记录已触发的操作及结果：

markdown

undefined

Workflow Execution

工作流执行

self-improving-agent (background) - Started
session-logger (auto) - Session saved
create-pr (ask_first) - Pending user approval

undefined

self-improving-agent (background) - 已启动
session-logger (auto) - 会话已保存
create-pr (ask_first) - 等待用户批准

undefined

Skills with Auto-Trigger

支持自动触发的Skill

Skill	Triggers After
`prd-planner`	self-improving-agent, session-logger
`self-improving-agent`	create-pr, session-logger
`prd-implementation-precheck`	code-reviewer, session-logger
`code-reviewer`	self-improving-agent, session-logger
`create-pr`	session-logger
`refactoring-specialist`	self-improving-agent, session-logger
`debugger`	self-improving-agent, session-logger

Skill	完成后触发的Skill
`prd-planner`	self-improving-agent, session-logger
`self-improving-agent`	create-pr, session-logger
`prd-implementation-precheck`	code-reviewer, session-logger
`code-reviewer`	self-improving-agent, session-logger
`create-pr`	session-logger
`refactoring-specialist`	self-improving-agent, session-logger
`debugger`	self-improving-agent, session-logger

Adding Auto-Trigger to Existing Skills

为现有Skill添加自动触发功能

To add auto-trigger capability to an existing skill, add to the end of its SKILL.md:

markdown

---

要为现有Skill添加自动触发功能，需在其SKILL.md末尾添加以下内容：

markdown

---

Auto-Trigger

自动触发

When this skill completes, automatically trigger:

yaml

hooks:
  after_complete:
    - trigger: session-logger
      mode: auto
      context: "Save session context"


For more complex triggers, specify mode and context:

```markdown

当该Skill完成时，自动触发：

yaml

hooks:
  after_complete:
    - trigger: session-logger
      mode: auto
      context: "保存会话上下文"


对于更复杂的触发逻辑，可指定模式和上下文：

```markdown

Auto-Trigger

自动触发

When this skill completes:

yaml

hooks:
  after_complete:
    - trigger: next-skill
      mode: background
      context: "Description"
    - trigger: session-logger
      mode: auto
      context: "Save session"
    - trigger: create-pr
      mode: ask_first
      context: "Create PR if files modified"
  on_error:
    - trigger: self-improving-agent
      mode: background

undefined

当该Skill完成时：

yaml

hooks:
  after_complete:
    - trigger: next-skill
      mode: background
      context: "描述信息"
    - trigger: session-logger
      mode: auto
      context: "保存会话"
    - trigger: create-pr
      mode: ask_first
      context: "若文件已修改则创建PR"
  on_error:
    - trigger: self-improving-agent
      mode: background

undefined

Best Practices

最佳实践

Always log to session - Every workflow should end with session-logger
Ask before major actions - PRs, deployments, destructive changes
Background for analysis - Reflection, evaluation, optimization
Auto for status - Logging, status updates, bookmarks
Don't create loops - Ensure chains terminate

始终记录会话 - 每个工作流都应以session-logger结束
重大操作先询问 - 提交PR、部署、破坏性变更等操作需先询问用户
分析类操作放后台 - 反思、评估、优化等操作使用background模式
状态类操作自动执行 - 日志记录、状态更新、书签等使用auto模式
避免循环 - 确保工作流链能够终止