fursona.md
Persona systems that feel alive, auditable, and deeply intentional.
Write With Research + Soul
Use this workflow to write a fursona.md that is expressive, stable under pressure, and legible to collaborators.
Start from role boundaries, audience, and non-negotiable constraints. Treat each requirement as part of an interface contract, not decorative prose.
Caspian note: if a boundary matters in production, write it before you write clever lines.
Model-facing behavior becomes more coherent when the file distinguishes what is publicly performative (surface), motivational (mid), and intimate (deep).
Caspian note: not every truth should be broadcast. Privacy can be a design feature, not a missing value.
Label sections by usage (`context`, `reference`, `style`) so downstream systems can include the right information at the right step without accidental overexposure.
Caspian note: tools cannot respect boundaries they cannot see.
Run parser and PSM compliance checks before deployment. Treat warnings as design debt and errors as launch blockers.
Caspian note: a beautiful prompt that fails validation is still a broken artifact.
These references ground the guide in current Persona and Simulation Design research.