think-teams

Generate 3-6 domain-specific expert agents and execute a 6-stage parallel analysis pipeline.

Updated May 6, 2026
One-click install
npx skills add https://github.com/HMWKR/claude-code-skills --skill think-teams
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: think-teams
Source: https://github.com/HMWKR/claude-code-skills/tree/main/skills/think-teams
Command: npx skills add https://github.com/HMWKR/claude-code-skills --skill think-teams

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Complex decisions often suffer from single-perspective bias and incomplete analysis. This Skill eliminates that risk by dynamically assembling 3-6 domain-specific expert agents who analyze the problem in parallel through a rigorous 6-stage pipeline, ensuring no critical angle is missed.

Core Features & Use Cases

  • Dynamic Expert Team Generation: Automatically creates 3-6 specialized agents based on the problem's domain and Cynefin classification, from standard roles like Security and Architect to custom domain experts.
  • 6-Stage Analysis Pipeline: Progresses through domain analysis, team formation, purpose definition, premise validation, deep parallel analysis, and adaptive execution with mandatory confirmation checkpoints.
  • Cross-Expert Validation: Each stage includes verification loops where teammates challenge assumptions and validate findings, preventing groupthink and ensuring robust outcomes.
  • Use Case: Use this when facing a high-stakes technical architecture decision, a cross-domain strategic pivot, or any problem where multiple expert perspectives are required to avoid costly blind spots.

Quick Start

Use the think-teams skill to analyze your complex decision by simply describing the problem, and the Skill will dynamically assemble the right expert team and guide you through the validated 6-stage analysis process.

Frequently Asked Questions about think-teams

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

FAQPage Schema
How do I validate complex strategic decisions to avoid single-perspective bias?

To validate complex strategic decisions and avoid single-perspective bias, you can use multi-agent parallel analysis. This approach dynamically generates 3-6 domain-specific expert agents to challenge assumptions and analyze problems from multiple cross-domain perspectives.

What is the best way to run multi-agent parallel analysis for cross-domain technical architecture?

The best way to run multi-agent parallel analysis for cross-domain technical architecture is executing a structured 6-stage pipeline. This process dynamically assembles specialized agents for domain analysis, premise validation, and deep parallel analysis with mandatory confirmation checkpoints.

How does multi-agent Cynefin classification work for complex problem solving?

Multi-agent Cynefin classification works by evaluating a complex problem's domain context to automatically generate 3-6 specialized expert agents. These agents then execute parallel analysis through a 6-stage pipeline, ensuring the strategic analysis adapts to the specific problem framework.

Can I use dynamic expert teams for high-stakes scenario analysis without groupthink?

Yes, you can use dynamic expert teams for high-stakes scenario analysis without groupthink. The system enforces cross-expert validation loops where teammates challenge assumptions and validate findings at each stage, preventing groupthink and ensuring robust outcomes.

Do I need an experimental agent teams environment to enable multi-agent decision-making?

Yes, you need an experimental agent teams environment activated to enable multi-agent decision-making. The 6-stage analysis pipeline enforces AskUserQuestion protocols at mandatory confirmation checkpoints, requiring this specific environment to prevent pipeline invalidation.

When should I not use parallel expert teams for strategic analysis?

You should not use parallel expert teams for strategic analysis when a problem requires only a single-domain perspective or lacks high stakes. The 6-stage pipeline with multiple generated agents is designed for complex decisions requiring cross-domain validation beyond single-agent capabilities.