research-and-go

Decompose broad goals into sub-plans and execute them sequentially.

Updated Apr 2, 2026
One-click install
npx skills add https://github.com/zillabunny/GoblinVolleyball-1.1 --skill research-and-go-zillabunny
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: research-and-go
Source: https://github.com/zillabunny/GoblinVolleyball-1.1/tree/main/.claude/skills/research-and-go
Command: npx skills add https://github.com/zillabunny/GoblinVolleyball-1.1 --skill research-and-go-zillabunny

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates end-to-end autonomous goal planning and execution, enabling teams to turn broad objectives into concrete, executable plans without manual handoffs.

Core Features & Use Cases

  • Decomposes broad goals into focused sub-plans and drafts each with adversarial review to improve robustness.
  • Generates a meta-plan and executes all phases sequentially, with automated verification and reporting.
  • Use Case: For complex product development, define a high-level goal and let the system plan, draft, and run the work without intermediate approvals.

Quick Start

Invoke /research-and-go <description> to have Claude decompose your goal, draft sub-plans with adversarial review, and execute them end-to-end.

Frequently Asked Questions about research-and-go

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

FAQPage Schema
How do I automate end-to-end autonomous planning for complex project goals?▼

End-to-end autonomous planning is automated by decomposing a broad objective into focused sub-plans, drafting each with adversarial review, and executing all phases sequentially without manual handoffs or intermediate user approvals.

What is adversarial plan drafting and how does it improve goal decomposition?▼

Adversarial plan drafting is a mechanism that reviews decomposed sub-plans to improve their robustness. It critically evaluates each drafted phase during goal decomposition, ensuring the final meta-plan can withstand edge cases before sequential execution begins.

Can I execute complex product development tasks without intermediate user approvals?▼

Yes, you can execute complex product development tasks without intermediate approvals by generating a meta-plan from decomposed sub-plans and running all phases sequentially. The system automates verification and reporting, requiring no user confirmation during end-to-end execution.

What's the best way to turn broad objectives into executable sub-plans automatically?▼

The best way to turn broad objectives into executable sub-plans is using an autonomous planning system that applies goal decomposition and adversarial review. It drafts focused sub-plans, generates a meta-plan, and handles sequential execution end-to-end.

Does autonomous goal execution require manual verification between sub-plan phases?▼

Autonomous goal execution does not require manual verification between phases. Once the meta-plan is generated from adversarially reviewed sub-plans, the system executes all phases sequentially with automated verification and produces a post-run report.

When do I need automated meta-plan generation for complex project workflows?▼

You need automated meta-plan generation when complex project workflows require decomposing broad goals into multiple sub-plans and executing them sequentially end-to-end. It eliminates manual handoffs by orchestrating autonomous planning, drafting, and execution without intermediate approvals.