Persona

Not everyone on your team is a developer. The CEO reviews PRs. The designer checks implementations. The PM asks for features. Each person has a different background, and the AI should talk to them differently.

hub persona creates a personal profile stored locally (.hub/persona.yaml) that tells the AI who it’s talking to. When you run hub generate, the persona is injected into AGENTS.md so the agent adapts its communication style automatically.

Quick Start

hub persona

This launches an interactive TUI that asks:

1. Your name
2. Your role (CEO, Designer, Backend Dev, QA Engineer, etc.)
3. Your technical level (non-technical, beginner, intermediate, advanced)
4. Any extra context about you (optional)
5. Your preferred language (English, Português, etc.)

Then run hub generate to apply it.

What Changes

The persona injects a ## User Persona section into AGENTS.md with specific communication instructions. Here’s what each level does:

Non-technical

For CEOs, PMs, designers, and anyone who doesn’t code:

• No jargon or acronyms without plain-language explanations
• Decisions framed as business impact, not implementation details
• Progress reported as product changes, not code changes
• No code snippets unless explicitly asked

Beginner

For people learning to code or new to the stack:

• Technical concepts explained briefly when first mentioned
• Code shown with context about what it does
• Encouraging, patient tone

Intermediate

For developers comfortable with code but still learning some tools:

• Normal technical language with explanations for niche concepts
• Focus on the “why” behind architectural choices

Advanced

For experienced developers:

• Concise, direct, technical communication
• Focus on trade-offs and edge cases
• No hand-holding

How It Works

The persona file is stored at .hub/persona.yaml — which is already gitignored (.hub/ is in the default .gitignore). Each team member has their own file on their machine.

name: João
role: CEO
technical_level: non-technical
context: "I focus on business metrics and product strategy"
language: Português

When hub generate runs, it reads this file and appends a persona section to AGENTS.md. The section includes specific instructions for the AI about how to communicate with that person.

Example Output

For a non-technical CEO, the generated section looks like:

## User Persona

You are talking to **João**, who is a **CEO**.

João is not technical. Adapt your communication:
- Never use jargon, acronyms, or technical terms without explaining them first.
- Explain decisions in terms of business impact, not implementation details.
- When showing progress, focus on what changed for the user/product.
- Keep responses short and focused on what matters to them.
- Never show code snippets unless explicitly asked.

Additional context about João: I focus on business metrics and product strategy

Always communicate with João in **Português**.

Per-Person, Not Per-Team

The hub.yaml is shared across the team. The persona is personal — it lives in .hub/ which is gitignored. Each person runs hub persona once on their machine, then hub generate picks it up automatically.

This means the same workspace can serve a CEO who wants business summaries and a senior dev who wants raw technical details.

Updating Your Persona

Run hub persona again — it pre-fills your existing answers so you can just change what you need.