exploration

Generates distinct decision options with evidence, unknowns, and assumption challenges.

4|2|Updated Jan 27, 2026
One-click install
npx skills add https://github.com/Arete-Consortium/ai-skills --skill exploration-arete-consortium
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: exploration
Source: https://github.com/Arete-Consortium/ai-skills/tree/main/workflows/exploration
Command: npx skills add https://github.com/Arete-Consortium/ai-skills --skill exploration-arete-consortium

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When facing a decision with an unmapped option space, teams often jump to a single solution too early or produce shallow brainstorms. This Skill scaffolds divergent thinking by generating genuinely distinct alternatives, surfacing unknowns, and challenging assumptions before any commitment is made. ## Core Features & Use Cases - Structured Option Generation: Produces an option table with id, name, premise, cheapest proof, and biggest risk for each alternative. - Unknown and Assumption Surfacing: Identifies the three highest-leverage unknowns with cheap tests and three assumptions that may not hold. - Failure-Condition Self-Checks: Detects duplicate options, vague "it depends" reasoning, premature recommendations, and missing unknowns before emitting output. - Use Case: A team deciding how to scale their API invokes /exploration with their objective, constraints, and assumptions, and receives a table of distinct architectural options with cheap validation tests instead of a premature recommendation. ## Quick Start Ask Claude to run /exploration with your decision objective, known constraints, current assumptions, evidence base, and excluded paths to receive a structured option table with unknowns and assumption challenges.

Frequently Asked Questions about exploration

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

FAQPage Schema
How do I generate distinct options for an unmapped decision?▼

Invoke /exploration with your objective, context, assumptions, evidence base, and exclusions. The skill produces an option table with premise, cheapest proof, and biggest risk per option, plus the three highest-leverage unknowns with cheap tests.

What is the difference between exploration mode and evaluation mode?▼

Exploration generates and names distinct options without ranking or recommending a winner. Evaluation ranks and decides among options. Run exploration first, then invoke /evaluation on its output when you need a pick.

What inputs does the exploration skill require?▼

It requires five inputs: objective (what you are deciding), context (known facts and constraints), assumptions (current beliefs to challenge), evidence base (docs, data, prior decisions), and exclusions (rejected or out-of-scope paths). Missing inputs are requested before proceeding.

When should I not use exploration mode?▼

Do not use it when the option space is already mapped and you need a recommendation, or when you are in build mode executing a known plan. Exploration is divergent thinking; it clears any active build goal loop on entry.

How does exploration mode prevent low-quality brainstorm output?▼

It self-checks failure conditions before emitting: collapsing duplicate options, rejecting vague "it depends" statements, blocking premature recommendations, and flagging sessions that surface no unknowns as suspicious.