ux-visual-evaluation
Skill2 starsChanged 9 months agoMITView on GitHub
Use after visual-feedback-loop confirms basic correctness, or when UI changes are complete and a screenshot is available, to evaluate screenshots for visual design quality — hierarchy, whitespace, affordances, and consistency
In this repository11 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- architecture-decision-recordsSkill
- boy-scout-ruleSkill
- domain-driven-designSkill
- functional-core-imperative-shellSkill
- git-commit-messageSkill
- react-best-practicesSkill
- revise-for-claritySkill
- ux-design-principlesSkill
- ux-prototypingSkill
- visual-feedback-loopSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
After visual-feedback-loop: When visual-feedback-loop's 5-item check passes.
- Distinctive: line 14
Independent: When UI changes are complete and a screenshot is available, even if visual-feedback-loop was not run.
- Distinctive: line 16
If the task is non-UI and no screenshot is available → skip.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
UX Visual Evaluation — Does It Look Good and Feel Usable?
Second pass after visual-feedback-loop confirms the UI renders correctly. Same screenshot, deeper evaluation. Drawn from Refactoring UI and Nielsen's usability heuristics.
When to Run
After visual-feedback-loop: When visual-feedback-loop's 5-item check passes. Uses the existing screenshot — don't take a new one.
Independent: When UI changes are complete and a screenshot is available, even if visual-feedback-loop was not run. Take or request a screenshot, then evaluate.
If the task is non-UI and no screenshot is available → skip.
Implement → Tests Pass → basic visual check (visual-feedback-loop) → UX evaluation (this skill) → RefactorVisual Design Criteria
Evaluate the screenshot for:
- Visual hierarchy — Does the most important element stand out? Primary actions: high-contrast solid buttons. Secondary actions: outline or lower contrast. Supporting text: lighter color, not just smaller. If everything looks equally prominent, nothing guides the user's eye.
- Whitespace — Generous breathing room? More space between groups than within them. Use whitespace to create separation instead of borders. Too much space is better than too little.
- Spacing consistency — Do margins, padding, and gaps follow a consistent scale? Look for one section with tight padding next to another with loose padding. Same-level elements should have the same spacing.
- Color usage — Is color strategic, not decorative? Primary actions should have the strongest color. Status indicators (success, error, warning) should use distinct, conventional colors. Text on colored backgrounds needs sufficient contrast.
- Borders vs. alternatives — Are borders overused? Prefer background color changes, subtle shadows, or increased whitespace. Borders make interfaces look busy.
Usability Criteria
- Affordances — Do interactive elements look interactive? Buttons should look like buttons, links like links. If you can't tell what's clickable from the screenshot, neither can the user.
- System status — Does the UI show where the user is and what's happening? Look for: highlighted nav items, empty state guidance, loading indicators, selected/active states.
- Consistency — Does this screen match the rest of the app? Same button styles, heading sizes, spacing patterns. New components should feel like they belong alongside existing ones.
Iteration
Same as visual-feedback-loop: fix, confirm test passes, re-evaluate. 3-attempt cap per issue — if unresolved, note as follow-up and proceed to refactoring.
What This Skill Does NOT Cover
- Whether the UI renders correctly —
visual-feedback-loop - Design decisions (who is the user, what's their goal) —
ux-design-principles - Accessibility audits —
react-best-practicesand dedicated tools
Related Skills
- Prerequisite:
visual-feedback-loop(basic correctness check before design evaluation),ux-design-principles(design rationale informs evaluation criteria) - Next:
boy-scout-rule(refactoring after visual verification is complete)
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

