Design of Everyday Things Skill: Score UX Clarity

What is the design-everyday-things skill?

This skill ends with a number. Its Quick Diagnostic has five rows, each worth two points: discoverability, evaluation, error recovery, mapping and constraints. A score in the 9 to 10 band means users act without instructions, understand every outcome and recover from any error.

Version 1.4.0 comes from the wondelai skills collection, which turns books into frameworks for agents. This one follows Don Norman's The Design of Everyday Things and applies it to software as well as physical products.

What changes in your agent's output

• Questions like why a screen is confusing get answered with named causes, such as a missing signifier, a poor mapping or absent feedback.
• A reviewed screen receives a score out of 10, with bands that say whether one gulf or error path is broken.
• Error messages are checked against a five point list: what went wrong, how to fix it, no blame, work preserved and an alternative path.
• Flat designs are checked for lost perceived affordances, where flat design hides whether something is a button or a label.

The ideas it applies

• The Gulf of Execution and the Gulf of Evaluation: how do I do it, and what just happened?
• Seven principles: discoverability, affordances, signifiers, mappings, constraints, feedback and conceptual models.
• Affordance types: real, perceived, hidden, false and anti-affordance.
• Feedback timing: about 0.1 seconds feels instant, 1 second is noticeable, and 10 seconds needs a progress indicator.
• Slips versus mistakes, with a design fix listed for each error type.
• The seven stages of action, from goal to comparing the result with what you wanted.

Who the design-everyday-things skill is for

Best for

• Diagnosing a flow where users hesitate, misclick or blame themselves. critique pairs well when you want a broader written review.
• Form and checkout work where error prevention matters. See form-design for field-level detail.

Not for

• Visual styling decisions such as palette, type or spacing. The skill judges clarity and not looks.
• Platform rules for one operating system. Its README lists a separate iOS design skill in the same collection.
• Replacing user research. The skill itself says to test with real users, noting that 5 reveal about 85 percent of problems.

Use it with UiChemy

The skill judges screens, whatever tool produced them. Review a page before it goes live, then fix the issues in UiChemy Composer, which edits WordPress pages visually in Elementor or Gutenberg. Click Get Started. For another review lens, try interface-review.

How to install Design of Everyday Things Skill: Score UX Clarity

/plugin marketplace add wondelai/skills

1. Add the marketplace

   In Claude Code, run /plugin marketplace add wondelai/skills.

2. Install a collection that includes it

   Run /plugin install ux-design@wondelai-skills. The README lists Design of Everyday Things in that collection, and also in product-innovation@wondelai-skills. These install the whole collection, not this skill alone.

3. Other agents

   The README says the skills suit Codex, Cursor, OpenClaw, Hermes Agent and other agentskills.io agents. It also describes a Codex plugin marketplace file at .agents/plugins/marketplace.json, managed from the plugins menu in the Codex TUI.

Check that it loaded

Ask your agent why a sign up screen is confusing and to score it. With the skill loaded the reply should name signifiers, mappings or feedback and end with a score out of 10 from the five row diagnostic.

Read the official instructions on GitHub

Key features

• Five row diagnosticScore two points for each satisfied row: discoverability, evaluation, error recovery, mapping and constraints. The scoring goal is 10 out of 10.
• The two gulfsEvery interaction is split into the Gulf of Execution and the Gulf of Evaluation, each with the questions users ask and the tools that bridge it.
• Affordances and signifiersAffordances come in five types. Signifiers are deliberate, accidental or social, and the skill lists digital ones such as cursor change, icons with labels, and color plus position.
• Natural mappingsControls should sit near what they affect and mirror the layout of the content. The skill rates examples as natural or poor and names proximity, spatial, cultural and sequential techniques.
• Constraint typesPhysical, cultural, semantic and logical constraints, with digital forms such as input validation, disabled states, forced sequence and undo.
• Feedback thresholdsDirect manipulation needs a response within 0.1 seconds, a 1 second delay needs a changed cursor, and a 10 second wait needs a progress bar.
• Conceptual model checkCompares the design model, the user's model and the system image, with a thermostat example showing how a wrong user model leads to wrong actions.
• Slips versus mistakesSplits human error into slips, such as action slips and mode errors, and mistakes, such as rule-based and knowledge-based ones, each paired with a design fix.
• Seven stages of actionWalk any interaction from goal to comparison and find the stage where users get stuck.
• Ten reference filesDeeper files cover affordances, case studies, conceptual models, constraints, feedback, human error, mappings, seven stages, signifiers and two gulfs.

Frequently asked questions