Design Review

Most AI design feedback guesses: it flags issues inconsistently, pads the list to look thorough, and critiques a text description instead of the actual pixels. Do the opposite — a fixed rubric applied to what you actually see.

Full rubric + checklists are in the SaglitzDesign MCP (npx saglitzdesign-mcp): docs design-critique-scoring, accessibility, principles-heuristics; tools design_review_checklist, plus the deterministic auditors that turn opinions into numbers — design_lint, audit_design_system, audit_accessibility, audit_ux_copy — and the /critique_screenshot and /design_review workflows.

Method

1. Look first, judge second. Describe what you actually see — layout, hierarchy, the primary action, which states are shown. If you're unsure what an element is, say so; don't invent.
2. Apply the fixed rubric. Score each of Nielsen's 10 heuristics 0–4 (total /40). Use the same rubric every time so scores are reproducible. Bands: 34–40 excellent, 26–33 solid, 18–25 needs work, <18 rework.
3. Cite specific elements. Every finding points to a concrete element: "the secondary 'Learn more' button competes with the primary CTA — two filled buttons," not "improve hierarchy."
4. No padding. Report only real issues. If the screen is genuinely good, a short list is the correct answer.
5. Rank by severity P0→P3, each with one concrete fix and the rule it comes from.

What to check (fast pass)