skills.rest
WenyuChiou
WenyuChiouCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-task-splitter

Split a high-level goal into a DAG of subtasks and emit .coord/plan.yml plus per-agent task files.

23|5|Updated Apr 28, 2026
Search Tags:#planning#codex#claude#multi-agent#gemini#dag#task-decomposition
One-click install
npx skills add https://github.com/WenyuChiou/agent-collab-skills --skill agent-task-splitter
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-task-splitter
Source: https://github.com/WenyuChiou/agent-collab-skills/tree/main/skills/agent-task-splitter
Command: npx skills add https://github.com/WenyuChiou/agent-collab-skills --skill agent-task-splitter

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Bridge between a high-level goal and the multi-agent execution pipeline. You write .coord/plan.yml (the DAG) and the per-agent task files. The delegate skills codex-delegate and gemini-delegate invoke the agents using those task files. The plan stops at file generation — this skill does not itself run the agents.

Core Features & Use Cases

  • Decompose a high-level goal into a DAG of subtasks and classify each by agent (codex / gemini / claude).
  • Emit a .coord/plan.yml plus per-agent task files for Codex, Gemini, and Claude.
  • Enable downstream delegate skills to invoke agents and reconcile outputs.

Quick Start

Trigger the splitter by providing a high-level goal and optional constraints; it will output a round-1 plan and per-agent task files ready for downstream delegate skills.

Frequently Asked Questions about agent-task-splitter

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I split a high-level goal into a DAG of subtasks for multi-agent runs?▼

To split a high-level goal into a DAG of subtasks, you provide the goal and optional constraints to generate a .coord/plan.yml file containing 2-7 subtasks with defined dependencies and verifiable success criteria.

How does multi-agent planning handle task classification for Codex, Gemini, and Claude?▼

Multi-agent planning classifies each subtask in the DAG by assigning it to a specific agent, automatically generating per-agent task files tailored for Codex, Gemini, and Claude to enable parallel execution.

Can I use this task decomposition approach to directly execute Codex and Gemini agents?▼

No, this task decomposition approach stops at file generation and does not directly execute agents. You need downstream delegate skills to invoke the agents using the generated per-agent task files.

What is the best way to structure dependencies in a multi-agent subtask plan?▼

The best way to structure dependencies in a multi-agent subtask plan is by defining a Directed Acyclic Graph (DAG) within the .coord/plan.yml schema, ensuring proper dependency handling and verifiable success criteria.

Does multi-agent task decomposition require a specific file format for plan reconciliation?▼

Yes, multi-agent task decomposition requires the .coord/plan.yml schema and specific per-agent task file formats for Codex, Gemini, and Claude to produce a reconciliation-ready plan.

What are the limitations of generating a DAG plan for multi-agent execution?▼

A limitation of generating a DAG plan is that the splitter creates a maximum of 2-7 subtasks and only emits files, meaning it cannot independently run agents or manage runtime execution errors.

Related Skills

Explore More Software Engineering Skills→
OzeroHAXOzeroHAX
32

project-fast-proto-spec

...

Community
Intermediate
Goblin1024Goblin1024
6

frontend-ui-ux

Route frontend UI/UX design tasks to the designer agent or Gemini MCP.

Community
Intermediate
VetCodersVetCoders
1

vc-delegate

Delegate bounded tasks to Claude subagents via the Task tool.

Official
Intermediate
WenyuChiouWenyuChiou
23

agent-acceptance-gate

Evaluate per-task success criteria and reconciler input to produce PASS/FAIL verdicts.

Community
Advanced