How does /explore differ from /research?

/explore is a divergent phase where multiple agents debate various options and risks, whereas /research is a convergent, single-agent process meant to formalize specific technical specs.

When should I use the Explore skill?

Use it at the start of a feature when the design space is wide, there are multiple viable approaches, or you need to understand both codebase constraints and external options simultaneously.

What kind of output does the Explore skill produce?

The primary output is an 'exploration-brief.md' file which includes team findings, debate outcomes, trade-offs, and a recommended technical approach.

Can I define custom roles for the exploration team?

Yes, you can provide custom roles in the arguments, such as 'security specialist' or 'database architect', to tailor the exploration to your specific needs.

Multi-Agent Problem Explorer

Name: Multi-Agent Problem Explorer
Author: AI-Builder-Team

byAI-Builder-Team

•

Productivity & Workflow

Conducts divergent, multi-perspective exploration of complex problem spaces by spawning and coordinating specialized AI agent teams.

The Explore skill transforms Claude Code into a project lead that manages a team of specialized agents to investigate complex technical problems. By assigning distinct roles—such as a Codebase Analyst, an External Researcher, and a Critic—the skill facilitates parallel discovery and constructive debate. This multi-perspective approach surfaces hidden risks, evaluates architectural tradeoffs, and challenges assumptions before formal specifications are written. The process culminates in a comprehensive exploration brief that serves as a high-quality foundation for the subsequent research and implementation phases of the development lifecycle.

Key Features

01Automated Role Assignment (Analyst, Researcher, Critic)

022 GitHub stars

03Parallel Multi-Agent Investigation

04Customizable Team Roles for Niche Domains

05Constructive Agent-to-Agent Debating

06Comprehensive Exploration Brief Generation

Use Cases

01Evaluating architectural tradeoffs for new features with a wide design space

02Investigating complex problem domains requiring both codebase analysis and external research

03Stress-testing proposed solutions via an automated 'Devil's Advocate' agent

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ai-builder-team/ai-builder-plugin-marketplace explore

For use in Claude.ai and ChatGPT

Download Skill