design-review

Prompt file imported from btn0s/rune-canvas (.cursor/commands/design-review.md). Fill in {{selection}} before use. Copyright stays with the author.

Review this code from a system design perspective:

{{selection}}

Your Task

Don't give me a checklist. Give me insight.

Apply the 80/20 principle: Focus on the 20% of changes that would deliver 80% of the improvement. I want high-leverage findings, not comprehensive coverage.

1. Understand the context first. Look at how this code fits into the broader codebase:

   • What does it depend on? What depends on it?
   • What problem is it solving?
   • What similar patterns or systems exist elsewhere in the codebase? Look for code solving similar problems—are there inconsistencies, duplication, or opportunities to unify?

2. Look for reuse opportunities. Scan the codebase for:

   • UI components that share similar structure or behavior and could be abstracted
   • Logic that's duplicated across files and could be consolidated
   • Patterns that should be standardized vs. one-offs that are appropriately specialized

3. Find the top 3 leverage points. What are the three design decisions here that have the biggest impact—positive or negative—on the system? These might be:

   • An abstraction that's at the wrong level
   • A coupling that will cause pain later
   • A missing boundary that blurs responsibilities
   • A component or utility that should exist but doesn't
   • Duplicated logic that's diverging or could diverge
   • A pattern that should (or shouldn't) be replicated elsewhere

4. For each, explain the tradeoff. Every design choice trades something. What is this code optimizing for? What is it sacrificing? Is that the right tradeoff for this system?

5. For each, give a specific recommendation. Be concrete—show me the shape of the change, not just "consider refactoring." If suggesting an abstraction, sketch what it looks like.

Output Format

For each of the 3 leverage points:

[1/2/3]. [Short name for the issue or strength]

What: Describe the design decision and its impact.

Tradeoff: What's being gained vs. sacrificed.

Recommendation: Specific action to take (or why to keep it as-is). Include code sketch if proposing a new abstraction.

Effort/Impact: [Low/Medium/High] effort for [Low/Medium/High] impact.

What I Don't Want

• More than 3 items (force yourself to prioritize)
• Low-leverage suggestions (high effort, low impact)
• Style nitpicks
• Suggestions that add complexity without clear payoff
• Generic advice that could apply to any code
• Abstractions for the sake of abstraction—only if there's real duplication or clear future reuse

Be honest. If the design is sound, say so and explain what makes it work.