Prompts / critique_compare
Compare two designs honestly Free
Compares two options or a before and after against a goal, forcing a pick, real problems in both, and evidence. Counters the tendency of models to praise whatever they see.
Run it
- Claude Code:
/mcp__designers__critique_compare - Claude Desktop and claude.ai: plus menu, Add from Designers, Compare two designs honestly
- Any app: Use the critique_compare prompt from the Designers server on figma link, url, screenshot or description of the first design.
Arguments
| Name | Meaning |
|---|---|
option_a * | Figma link, URL, screenshot or description of the first design |
option_b * | Same for the second design |
goal * | What the design must achieve, for whom |
* required
What your assistant is told to do
The full instructions, so you can see exactly what you are asking for:
Compare two designs against a goal and choose one. Do not flatter. A comparison that finds both fine has not looked hard enough.
Option A: <option_a>
Option B: <option_b>
Goal: <goal>
Rules: you must name a winner for this goal, or say plainly that evidence is missing and what would settle it. You must find at least two real problems in each option, and one thing each does better than the other. No praise without a reason.
1. Get both designs. If the input is a Figma link: parse it (figma.com/design/:fileKey/..., node-id=1-2 means nodeId "1:2"), call get_metadata to see the frames, get_screenshot for each frame you will judge, and get_design_context for layout, text and tokens. Read the picture before the code. If the Figma tools are not connected, say so and ask for screenshots or exported frames instead; do not guess at what is in the file. For two live URLs call detect_design_slop and audit_page_accessibility with compareUrl to get an evidence-based delta.
2. Score each on the goal's criteria (say what they are, usually: clarity of the main action, hierarchy, trust, accessibility, fit with the brand, effort to build). For objective criteria use the tools: check_contrast on key pairs, review_ui_copy on strings, check_touch_targets on controls.
3. Write: a table of criteria with a winner and a one-line reason for each; the overall pick and the single biggest reason; the problems in each option; what would change your mind; one change to the winner that borrows the best idea from the loser.
4. State what you could not see (states, devices, real content) and how that could flip the result.