Human Stars

The Problem

When a question needs cross-domain judgment, a raw model answer can feel averaged: it speaks like a product manager, a philosopher, and an engineer at once, without a clear bias. Human Stars treats people as routing keys, so the reply carries an expected mental model, writing style, and value framing.

How It Works

The skill uses 42 built-in figures and natural-language signals to load different workflows:
- One-to-one dialogue: invoke a person, such as Buddha or Jobs, and receive a response framed by that persona
- Group collision: ask several people the same question and compare mental models
- Thinking-model toolbox: call first principles, moat, or combine multiple models
- Debate arena: pit two figures against each other and get a three-axis score for logic, creativity, and action
- Socratic challenge: receive consecutive counter-questions against a stated belief
- Scenario routing and star refinement: recommend figures by state, or add a new figure through a six-layer template

Routing prioritizes explicit signals: a name plus a question goes to dialogue; simultaneously or together goes to collision; debate or PK goes to arena; analyze with XX goes to the toolbox; vague states such as insomnia or confusion go to scenario recommendation. You can also bookmark lines and export the conversation.

Boundaries

It is useful for decision rehearsal, perspective switching, writing material, and thinking training. It should not replace clinical, legal, or crisis support. Persona replies are model-based roleplay, not the person's actual views; refined figures need template validation and user confirmation before being treated as reliable references.