create-brief

name	create-brief
description	Create a product brief through guided discovery. Asks questions to understand the product vision, users, constraints, and success criteria. Writes to .factory/specs/product-brief.md.
allowed-tools	Read, Write, Edit, Bash, AskUserQuestion

Path Resolution (Mandatory)

Before writing any artifact, resolve the canonical path via plugins/vsdd-factory/config/artifact-path-registry.yaml. Do not invent paths.

Read the registry at the start of this skill's procedure and verify the target path matches a registered pattern before calling Write. If the artifact type is not in the registry, use /vsdd-factory:register-artifact to add it first.

Hard Gate

Do NOT skip to PRD creation or architecture design. Every discovery section MUST be explored with the human. Do not auto-fill sections from assumptions.

Create Product Brief

Guide the user through creating a product brief — the L1 foundation that everything else builds on.

Templates

Read and follow the output format in:

• ${CLAUDE_PLUGIN_ROOT}/templates/product-brief-template.md — L1 product brief structure

Before Starting

1. Run /factory-health to ensure .factory/ is healthy.
2. Check if .factory/specs/product-brief.md already exists. If so, ask: update existing or start fresh?
3. Check .factory/specs/research/RESEARCH-INDEX.md for prior domain research. If research exists, read the relevant reports to inform the brief creation — don't ask questions the research already answers.

Discovery Process

Ask questions one at a time. Use multiple choice when possible.

1. Vision & Problem

• What problem does this product solve?
• Who has this problem? (target users)
• What happens if this problem isn't solved? (pain/stakes)

2. Users & Personas

• Who are the primary users?
• Who are secondary users?
• What's their technical sophistication?

3. Core Value Proposition

• What makes this different from alternatives?
• What's the one thing this must do well?
• What is explicitly out of scope?

4. Success Criteria

• How will you know this works?
• What metrics matter?
• What's the MVP vs full vision?

5. Constraints

• Technical constraints (language, platform, integrations)
• Timeline constraints
• Team/resource constraints
• Regulatory or compliance requirements

6. Prior Art

• Existing codebases to reference (for gene transfusion/semport)
• Competitor products
• Internal tools this replaces

Output

Write to .factory/specs/product-brief.md:

# Product Brief: <Product Name>

**Author:** <user name>
**Date:** <current date>
**Status:** draft

## Problem Statement
<What problem, who has it, what are the stakes>

## Target Users
<Primary and secondary personas>

## Value Proposition
<What makes this unique, core differentiator>

## Success Criteria
<Measurable outcomes>

## Scope
### In Scope
<What we're building>

### Out of Scope
<What we're explicitly not building>

## Constraints
<Technical, timeline, regulatory>

## Prior Art & References
<Existing codebases, competitors, reference implementations>

## Open Questions
<Things we need to resolve before moving to PRD>

Self-Review (before adversarial review)

Before routing to the next pipeline stage, check your own work:

1. Placeholder scan: Any "TBD", "TODO", incomplete sections, or vague requirements? Fix them now.
2. Internal consistency: Does the scope match the success criteria? Do constraints align with the value proposition?
3. Scope check: Is this brief focused enough for a single PRD, or does it describe multiple independent products?
4. Ambiguity check: Could any requirement be interpreted two different ways? Pick one and make it explicit.

Fix issues inline. This is a cheap filter — catch obvious gaps before the next stage.

After Writing

1. Commit to factory-artifacts: cd .factory && git add specs/product-brief.md && git commit -m "factory(phase-1): create product brief"
2. Invoke state-update to set phase to phase-1, status in-progress.
3. Tell the user: "Brief created. Next step: /create-prd to elaborate into a full PRD, or /create-domain-spec if the domain needs deeper modeling first."