personafy

Create evidence-based advisor persona skills from real people or archetypes.

10|11|Updated Feb 22, 2026
One-click install
npx skills add https://github.com/microsoft/amplifier-bundle-skills --skill personafy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: personafy
Source: https://github.com/microsoft/amplifier-bundle-skills/tree/main/skills/personafy
Command: npx skills add https://github.com/microsoft/amplifier-bundle-skills --skill personafy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you create a reusable advisor-persona skill from a real person or archetype, so an AI can review work through a consistent, opinionated lens instead of a generic one.

Core Features & Use Cases

  • Persona skill authoring: Draft a new sibling persona skill that fits an existing family and owns one distinct load-bearing question.
  • Evidence-grounded design: Mine real transcripts, corpora, or briefs for authentic voice, decision discipline, and verbatim steering quotes.
  • Steering validation: Prove the persona changes model behavior in a live session, then reduce it to the smallest effective form.
  • Use case: Use this when you want to turn a real reviewer style into a durable skill for code review, product critique, or advisory analysis.

Quick Start

Use the personafy skill to create an evidence-based persona skill for a reviewer modeled on a real subject.

Frequently Asked Questions about personafy

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

FAQPage Schema
How do I create an AI advisor persona that reviews work using a specific person's style?

To create an advisor persona, you build an evidence-based skill using real transcripts or corpora to capture an authentic voice and verbatim steering quotes. This grounds the AI in a consistent, opinionated lens rather than a generic response.

What is evidence-based persona skill authoring for LLM design?

Evidence-based persona skill authoring is the process of mining real-world text to extract a subject's decision discipline and voice. This data reduces the persona to a minimal reusable SKILL.md that consistently steers large language model behavior.

How do I validate that a reviewer lens skill actually changes model behavior?

You validate a reviewer lens by running the persona skill in a live session to prove it alters model behavior. Once the steering question is verified, the skill is reduced to its smallest effective form for durable reuse.

Can I use real transcripts to build a reusable AI persona for code review?

Yes, you can mine real transcripts and corpora to draft a sibling persona skill for code review. The skill defines a single distinct load-bearing question and requires family-template consistency to maintain an authentic, opinionated lens.

What's the best way to reduce a persona skill to a minimal reusable format?

The best way to reduce a persona skill is to validate its steering behavior in a live session first, then strip it down to the smallest effective SKILL.md. This ensures the behavior steering remains grounded and family-template consistent.