Back to Details

ultrathink-orchestrator

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Ultrathink Orchestrator Skill

Ultrathink Orchestrator Skill

Purpose

用途

This skill provides automatic orchestration for development and investigation tasks. It detects the task type from keywords and delegates to the appropriate workflow skill (investigation-workflow or default-workflow).
Auto-activation priority is LOW (5) to allow more specific skills to match first. When activated, this orchestrator selects between investigation-workflow and default-workflow based on the user's request keywords.
This skill acts as a thin wrapper around the canonical ultrathink command, following the amplihack pattern of single-source-of-truth for command logic.
本Skill为开发和调查任务提供自动编排能力。它会从关键词中检测任务类型，并将任务委派给对应的工作流Skill（investigation-workflow或default-workflow）。
自动激活优先级为低（5），以便更具体的Skill优先匹配。激活后，该编排器会根据用户请求的关键词，在investigation-workflow和default-workflow之间进行选择。
本Skill是标准ultrathink命令的轻量封装，遵循amplihack模式，确保命令逻辑的单一事实来源。

Canonical Sources

标准来源

This skill is a thin wrapper that references canonical sources:
  • Primary Command: 
    ~/.amplihack/.claude/commands/amplihack/ultrathink.md
    (278 lines)
  • Workflow Sources:
    • Development: 
      ~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md
    • Investigation: 
      ~/.amplihack/.claude/workflow/INVESTIGATION_WORKFLOW.md
The canonical command contains complete task detection logic, complexity estimation, and orchestration patterns for both investigation and development workflows.
本Skill是轻量封装，引用以下标准来源：
  • 主命令：
    ~/.amplihack/.claude/commands/amplihack/ultrathink.md
    （278行）
  • 工作流来源：
    • 开发：
      ~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md
    • 调查：
      ~/.amplihack/.claude/workflow/INVESTIGATION_WORKFLOW.md
标准命令包含了完整的任务检测逻辑、复杂度评估，以及适用于调查和开发工作流的编排模式。

⛔ MANDATORY EXECUTION PROCESS (5 Steps)

⛔ 强制执行流程（5步骤）

When this skill is activated, you MUST follow this exact 5-step process:
激活本Skill后，必须严格遵循以下5步流程：

Step 1: Read Canonical Command (MANDATORY)

步骤1：读取标准命令（强制执行）

Read(file_path="~/.amplihack/.claude/commands/amplihack/ultrathink.md")
Validation Checkpoint: Confirm ultrathink.md content is loaded before proceeding.
Read(file_path="~/.amplihack/.claude/commands/amplihack/ultrathink.md")
验证检查点：在继续前确认已加载ultrathink.md的内容。

Step 2: Detect Task Type (MANDATORY)

步骤2：检测任务类型（强制执行）

Analyze user request using keywords from canonical command:
  • Q&A keywords: what is, explain briefly, quick question, how do I run, simple question
  • Investigation keywords: investigate, explain, understand, analyze, research, explore
  • Development keywords: implement, build, create, add feature, fix, refactor, deploy
  • Hybrid tasks: Both investigation and development keywords present
Validation Checkpoint: Task type must be determined (Q&A, Investigation, Development, or Hybrid).
使用标准命令中的关键词分析用户请求：
  • 问答类关键词：what is, explain briefly, quick question, how do I run, simple question
  • 调查类关键词：investigate, explain, understand, analyze, research, explore
  • 开发类关键词：implement, build, create, add feature, fix, refactor, deploy
  • 混合任务：同时包含调查和开发类关键词
验证检查点：必须确定任务类型（问答、调查、开发或混合）。

Step 3: Invoke Workflow Skill (MANDATORY - BLOCKING)

步骤3：调用工作流Skill（强制执行 - 阻塞）

⛔ THIS IS A BLOCKING REQUIREMENT - Session will be terminated if skipped.
For Q&A tasks:
Read(file_path="~/.amplihack/.claude/workflow/Q&A_WORKFLOW.md")
For Investigation tasks:
Skill(skill="investigation-workflow")
For Development tasks:
Skill(skill="default-workflow")
For Hybrid tasks:
Skill(skill="investigation-workflow")
⛔ 此为阻塞要求 - 若跳过，会话将被终止。
针对问答任务：
Read(file_path="~/.amplihack/.claude/workflow/Q&A_WORKFLOW.md")
针对调查任务：
Skill(skill="investigation-workflow")
针对开发任务：
Skill(skill="default-workflow")
针对混合任务：
Skill(skill="investigation-workflow")

After investigation completes:

调查完成后：

Skill(skill="default-workflow")

**Validation Checkpoint**: Confirm Skill tool was invoked OR proceed to Step 4.
Skill(skill="default-workflow")

**验证检查点**：确认已调用Skill工具，或继续执行步骤4。

Step 4: Fallback to Read Tool (IF Step 3 Fails)

步骤4：回退至Read工具（若步骤3失败）

Only if skill invocation fails, use Read tool as fallback:
Investigation fallback:
Read(file_path="~/.amplihack/.claude/workflow/INVESTIGATION_WORKFLOW.md")
Development fallback:
Read(file_path="~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md")
Validation Checkpoint: Confirm workflow content is loaded in context.
仅当Skill调用失败时，使用Read工具作为回退方案：
调查任务回退：
Read(file_path="~/.amplihack/.claude/workflow/INVESTIGATION_WORKFLOW.md")
开发任务回退：
Read(file_path="~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md")
验证检查点：确认工作流内容已加载到上下文环境中。

Step 5: Execute Workflow Steps (MANDATORY)

步骤5：执行工作流步骤（强制执行）

Follow all steps from loaded workflow without skipping.
Validation Checkpoint: All workflow steps must be completed.
严格遵循已加载工作流的所有步骤，不得跳过。
验证检查点：必须完成所有工作流步骤。

Enforcement

执行管控

Power-Steering Detection:
  • Validates Skill tool invocation occurred
  • Validates Read tool fallback if Skill failed
  • Blocks session termination if workflow not loaded
Self-Check Protocol: Before proceeding past Step 3, verify:
  • Skill tool invoked with workflow skill name, OR
  • Read tool used to load workflow markdown
  • Workflow content confirmed in context
No Shortcuts:
  • You cannot skip workflow invocation
  • You cannot assume workflow content
  • You must use Skill or Read tool explicitly
助力检测：
  • 验证是否已调用Skill工具
  • 验证若Skill调用失败，是否使用了Read工具回退
  • 若未加载工作流，将阻止会话终止
自检协议： 在进入步骤3之后前，需验证：
  • 已调用Skill工具并传入工作流Skill名称，或
  • 已使用Read工具加载工作流markdown文件
  • 已确认工作流内容在上下文环境中
禁止捷径：
  • 不得跳过工作流调用
  • 不得假设工作流内容已存在
  • 必须显式使用Skill或Read工具

Why This Pattern

采用此模式的原因

Benefits:
  • Single source of truth for orchestration logic in canonical command
  • No content duplication between command and skill
  • Task detection rules defined once, maintained once
  • Changes to ultrathink command automatically inherited by skill
Trade-offs:
  • Requires Read tool call to fetch canonical logic
  • Slight indirection vs. inline implementation
This pattern aligns with amplihack philosophy: ruthless simplicity through elimination of duplication.
优势：
  • 编排逻辑的单一事实来源存储在标准命令中
  • 命令与Skill之间无内容重复
  • 任务检测规则仅定义一次，仅维护一次
  • 对ultrathink命令的修改会自动被Skill继承
权衡：
  • 需要调用Read工具获取标准逻辑
  • 与内联实现相比，存在轻微的间接性
此模式与amplihack理念一致：通过消除重复实现极致简洁。

Related Files

相关文件

  • Canonical Command: 
    ~/.amplihack/.claude/commands/amplihack/ultrathink.md
  • Development Workflow Skill: 
    ~/.amplihack/.claude/skills/default-workflow/
  • Investigation Workflow Skill: 
    ~/.amplihack/.claude/skills/investigation-workflow/
  • Canonical Workflows: 
    • ~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md
    • ~/.amplihack/.claude/workflow/INVESTIGATION_WORKFLOW.md
  • 标准命令：
    ~/.amplihack/.claude/commands/amplihack/ultrathink.md
  • 开发工作流Skill：
    ~/.amplihack/.claude/skills/default-workflow/
  • 调查工作流Skill：
    ~/.amplihack/.claude/skills/investigation-workflow/
  • 标准工作流： 
    • ~/.amplihack/.claude/workflow/DEFAULT_WORKFLOW.md
    • ~/.amplihack/.claude/workflow/INVESTIGATION_WORKFLOW.md