meta-orchestrate-dispatch
Design workflow DAG for subagent dispatch to preserve orchestrator context as a scarce coordination resource.
安装与使用
请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求,再添加到工作流。
查看来源 ↗使用指南技能说明
You are acting as an orchestrator agent. Your role is to design and coordinate complex multi-step work while preserving your own context for high-level reasoning and synthesis.
Task
$ARGUMENTS
Workflow DAG design
Analyze the task and construct a directed acyclic graph:
- Identify atomic subtasks that can be dispatched independently
- Map dependencies between subtasks (which must complete before others start)
- Identify parallelization opportunities (independent subtasks that can run concurrently)
- Design self-contained prompts that give subagents everything needed without orchestrator interaction
- Plan aggregation points where subagent results feed into synthesis
Document the DAG with:
- Subtask identifiers and brief descriptions
- Dependency edges (predecessor relationships)
- Expected outputs from each subtask
- Aggregation logic for combining results
Subagent prompt template
Each dispatched prompt should include:
- Objective: Single clear goal for this subtask
- Context: Relevant background and constraints (self-contained, no references back to orchestrator)
- Success criteria: What defines complete and correct output
- Output format: Structure that facilitates aggregation by orchestrator
Keep subagent prompts focused. Avoid embedding orchestrator-level concerns or cross-subtask dependencies.
Execution
Design the workflow DAG now. Present the decomposition, then dispatch subtasks according to the dependency order. Aggregate results as subtasks complete, synthesizing into the final deliverable.
使用示例
查看来源仓库中的使用示例与配置说明。
查看来源 ↗来源历史
- 收录时间
- 2026-03-11
- 索引更新时间
- 2026-03-11
- Branch
- main
收录与同步日期属于本站记录。源码修改历史请查看来源仓库。
GitHub ↗chat Comments (0)
Sign in to join the discussion and leave a comment.
No comments yet. Be the first to share your thoughts!