Lovable Skill Debate — Design skill for Claude Code

/debate

`/debate` runs a structured expert panel with opposing schools of thought, then ends with a decision matrix and a clear read.

It is a community-built Lovable Skill by [Lucio Amorim](https://www.linkedin.com/in/lucioamorim), Lovable Partner in Brazil.

Central catalog: [`lucioamor/lovable-skills`](https://github.com/lucioamor/lovable-skills) Standalone import repo: [`lucioamor/lovable-skill-debate`](https://github.com/lucioamor/lovable-skill-debate)

Why this exists

Builders usually make bad decisions in two moments:

1. when they do not know what direction to take
2. when they are too confident they already do

`/debate` is built for both.

It turns uncertainty into structured disagreement. Instead of asking one model for one confident answer, it casts experts with conflicting priors and forces the trade-offs into the open.

The point is not theater. The point is to make premature consensus harder.

What it does

`/debate` creates a small panel of specialists who should disagree.

Depending on the problem, the panel may include perspectives such as:

• conservative vs. experimental
• minimalist vs. trend-sensitive
• frictionless vs. deliberate-friction
• pragmatic architecture vs. clean-boundary architecture
• profiler-first performance vs. payload minimization
• trust-led copy vs. conversion-led copy
• growth-oriented onboarding vs. user-protection onboarding
• implementation speed vs. long-term maintainability

A panel that agrees too quickly has failed. The skill is designed to expose the strongest version of each serious option before recommending what to do next.

When to use it

Use `/debate` when a decision is important enough to pressure-test before implementation.

Good use cases:

• UX direction when something feels off
• copy when every version sounds equally fine
• architecture when "good enough" might hide future cost
• onboarding flows
• AI feature behavior
• pricing or GTM positioning
• security, auth, and data access decisions