design-criticlisted

Review a design and come back with the issues worth fixing, ranked, each with a fix. Use it when a screen, a flow, or a set of screens is ready for a hard look. It runs the right critique method from the UX + AI connector for what you have in front of you: a 30-second first-impression read, the full usability critique, a friction audit on a flow, a consistency check across screens, a structured accessibility pass, or the three objections most likely to block a design review. Outcome: issues ranked by what they cost users, each with a fix. Trigger on 'review this screen', 'critique this design', 'what is wrong with this', 'is this clear', 'why does this feel heavy', 'nobody clicks the button', 'these screens do not match', 'check accessibility', 'is this ready for design review', 'be honest about my design', or any request to review, audit, or pressure-test an interface.
IleanaMarcut/uxai-mcp-plugin · ★ 1 · Web & Frontend · score 64

Install: claude install-skill IleanaMarcut/uxai-mcp-plugin

# Design Critic ## What this is Six critique methods behind one door: a first-impression read, the full usability critique, a cognitive-load audit, a consistency check, an accessibility pass, and a pre-review objection hunt. The outcome is the same shape every time: issues ranked by what they cost users, each with a fix. ## When and why to use it Use it when a design exists and is ready for a hard look: before a design review, before shipping, or when something feels off and the feeling needs a name. What you have in front of you decides the method, because a critique run on the wrong scope produces confident findings about something you did not show. ## This skill runs the methods of the UX + AI MCP Each critique is a method served by the UX + AI connector. This file only routes; the methods do the work, and they improve on the connector without this file changing. ## How it works 1. **Check the connector.** The UX + AI MCP must be connected on a subscribed account. If it is unavailable, say so and stop. 2. **Look at what they have before choosing.** "Review my checkout" with one screenshot is a screen. The same words with five screens and a task is a flow. When the request and the material disagree, say which one you are working from. 3. **Fetch by id and follow the text exactly as returned.** These are the six methods this skill runs; never rebuild one from memory or from this file. ``` fetch(id: "quick-clarity-check") fetch(id: "heuristic-critique") f