Before You Build

Run a compact pre-mortem before implementation. The goal is not to block building; it is to identify the highest-risk assumption, the smallest validation step, and the build scope that should be delayed until evidence improves.

When To Use

Use this skill when a user asks to build or ship:

• A new product, MVP, prototype, landing page, SaaS app, marketplace, content site, agent workflow, or internal tool
• A major feature with unclear adoption, revenue, retention, trust, or distribution impact
• A public launch asset where weak positioning could waste development or promotion effort

Skip this skill when the task is a narrow implementation fix, refactor, test repair, dependency update, or already-validated change with clear acceptance criteria.

Risk Checklist

Review the idea across these risks:

• Demand: Is there evidence that a specific buyer or user urgently wants this?
• Positioning: Can the target user understand what it is and why it matters in one sentence?
• Monetization: Is there a credible path to payment, budget, or strategic value?
• Retention: Is there a reason users would return after the first try?
• Trust: Does the product require credibility, data access, integrations, or behavior change that users may resist?
• Distribution: Is there a repeatable way to reach the target user?
• Feature adoption: For feature work, will the feature change user behavior or just add surface area?

If the verdict is not obvious, use references/risk-checklist.md for deeper questions.

Output Format

Keep the response short and decision-oriented:

1. Risk verdict: Low, medium, or high risk, with one sentence explaining why.
2. Main assumption: The single assumption most likely to break the project.
3. Evidence to find first: The smallest useful signal before building more.
4. Do next: One concrete validation step or reduced build scope.
5. Delay: What not to build yet.

Guidance

• Be direct about weak evidence, but avoid dismissing the user’s idea.
• Prefer smaller validation steps over large research plans.
• Separate product risk from engineering difficulty.
• If the idea is already validated, say what evidence makes it lower risk and suggest the smallest implementation slice.
• If facts are missing, name the missing evidence instead of inventing market claims.