ux-heuristic-eval

Evaluates interfaces against Nielsen's 10 usability heuristics with severity-rated findings and remediation recommendations.

Updated Jul 2, 2026
One-click install
npx skills add https://github.com/geekatron/jerry-claude-plugin --skill ux-heuristic-eval-geekatron
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ux-heuristic-eval
Source: https://github.com/geekatron/jerry-claude-plugin/tree/main/skills/ux-heuristic-eval
Command: npx skills add https://github.com/geekatron/jerry-claude-plugin --skill ux-heuristic-eval-geekatron

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Teams without dedicated UX evaluators lack a structured way to audit interfaces for usability problems, leading to inconsistent reviews and unprioritized fixes. ## Core Features & Use Cases - Systematic Heuristic Inspection: Applies all 10 of Nielsen's usability heuristics sequentially to every screen or flow, producing findings rated on a 0-4 severity scale from Cosmetic to Catastrophic. - Remediation Roadmap: Generates actionable fix recommendations with Low/Medium/High effort estimates, deduplicated and ranked by severity. - AI-Interaction Supplements: Adds Transparency, Controllability, and Error Recovery heuristics when evaluating AI-driven product interfaces. - Use Case: Ask the parent /user-experience skill to audit a settings page; the orchestrator routes to this sub-skill, which produces a persisted evaluation report with ranked findings and handoff data for downstream Behavior Design and HEART Metrics analysis. ## Quick Start Ask the /user-experience skill to run a heuristic evaluation of your interface by providing screenshots or a Figma reference, for example: evaluate this dashboard design against Nielsen's heuristics.

Frequently Asked Questions about ux-heuristic-eval

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a heuristic evaluation on my interface?

Invoke the parent /user-experience skill with a request like 'evaluate this dashboard against Nielsen's heuristics' and provide screenshots or a Figma reference. The ux-orchestrator routes the request to the ux-heuristic-evaluator agent, which produces a severity-rated report.

What severity scale does a Nielsen heuristic evaluation use?

Findings are rated on Nielsen's 0-4 scale: 0 (not a problem), 1 (cosmetic), 2 (minor), 3 (major), and 4 (usability catastrophe). Only findings rated severity 2 or higher are propagated to downstream sub-skills like Behavior Design and HEART Metrics.

Can heuristic evaluation work without Figma access?

Yes, the evaluator operates in screenshot-input mode when the Figma MCP adapter is unavailable. The methodology is identical, but the report discloses degraded mode limitations such as inability to inspect component states or responsive behavior.

Does this skill evaluate AI product interfaces?

Yes, when the AI Product Flag is set, three supplementary heuristics are applied: AI-1 Transparency, AI-2 Controllability, and AI-3 Error Recovery. These findings are marked [AI-SUPPLEMENT] to distinguish them from Nielsen's core 10 heuristics.

What are the limitations of a single AI evaluator for usability audits?

A single evaluator typically finds only about 35% of usability problems compared to 75-80% with 3-5 human evaluators. For severity 3-4 findings, the skill recommends supplementing with at least one human reviewer before major design decisions.

When should I use behavior design instead of heuristic evaluation?

Use /ux-behavior-design (Fogg B=MAP) when diagnosing why users fail to take a specific action. Heuristic evaluation identifies design-level usability issues but does not trace behavioral root causes or measure quantitative UX metrics.