What is the heuristic-evaluation skill?
The heuristic-evaluation skill is a SKILL.md in the prototyping-testing plugin of Owl-Listener's designer-skills repository. It has an AI coding agent review an interface against Nielsen's ten heuristics, document each issue and rate severity from 0 to 4, without recruiting participants.
The heuristic-evaluation skill is part of the prototyping-testing plugin in Owl-Listener's designer-skills pack. It casts the agent as an expert in systematic heuristic evaluations of digital interfaces, finding problems before user testing.
Its description gives the trigger: you need findings without recruiting participants. For a facilitated team feedback session, it points to design-critique in the design-ops plugin, covered on the design critique page.
Nielsen's ten heuristics
- Visibility of system status.
- Match between system and real world.
- User control and freedom.
- Consistency and standards.
- Error prevention.
- Recognition over recall.
- Flexibility and efficiency.
- Aesthetic and minimalist design.
- Error recovery.
- Help and documentation.
Evaluation process and severity scale
The process defines scope, then walks through as a new user, as an experienced user and through each task flow. It documents each issue, rates severity and compiles prioritized findings. Severity runs from 0, not a problem, to 4, a catastrophe to fix before release.
Issue documentation
Every issue records the heuristic violated, a description, the location, a severity from 0 to 4, a screenshot or reference and a recommendation. The skill's issue documentation lists these six fields for each finding.
Who the heuristic-evaluation skill is for
Best for
- Getting expert findings on a screen or flow before any participants are recruited.
- Reviewers who want every issue tagged with the heuristic it violates and a severity. The UX heuristics skill also works from Nielsen's ten.
- Teams that pair the review with real user testing for a complete picture. The design review skill is another Nielsen based review.
Not for
- Facilitating a team feedback session. Its description names design-critique for that.
How to install Heuristic Evaluation Skill: Expert Review Without Participants
/plugin marketplace add Owl-Listener/designer-skillsAdd the marketplace
Enter the command above in your AI coding agent. It points the agent at the repository, and no plugin installs at this stage.
Install the plugin
Run
/plugin install prototyping-testing@designer-skills. prototyping-testing is the plugin that holds this skill, and the Discover tab under/pluginoffers the same choice.Gemini CLI route
Each plugin is also a Gemini CLI extension. Clone the repository to a temporary folder, create
.gemini/extensionsin your project root, and copy the repository's.gemini/extensionscontents into it.Name the scope
Tell the agent which screens or flows to evaluate and who the users are. It walks each flow and logs issues by heuristic.
Check that it loaded
Ask for a heuristic evaluation of one flow. Expect issues labeled with the heuristic violated and a severity from 0 to 4, followed by prioritized findings.
Key features
- Ten heuristicsVisibility of system status, match with the real world, user control and freedom, consistency and standards, error prevention, recognition over recall, flexibility and efficiency, minimalist design, error recovery, and help and documentation.
- Defined scopeThe first step sets which screens and flows are evaluated.
- Three walkthroughsWalk through as a new user, as an experienced user and through each task flow.
- Issue recordsEach issue notes the heuristic violated, a description, the location, severity, a screenshot or reference and a recommendation.
- 0 to 4 severity scale0 is not a problem, 1 cosmetic, 2 minor, 3 major and important to fix, 4 a catastrophe that must be fixed before release.
- Prioritized findingsThe final step compiles and prioritizes findings.
- Multiple evaluatorsThree to five evaluators are ideal, and they should evaluate independently before comparing.
- Real tasksFocus on real user tasks, not edge cases.
- Solutions includedSuggest solutions next to each problem, and combine the review with real user testing.
Frequently asked questions
What is the heuristic-evaluation skill?
It is a SKILL.md in the prototyping-testing plugin of Owl-Listener's designer-skills repository. It has an AI coding agent run an expert review of an interface against Nielsen's ten heuristics, document each issue and rate its severity, so problems surface before user testing.
How do I install heuristic-evaluation?
Run /plugin marketplace add Owl-Listener/designer-skills, then /plugin install prototyping-testing@designer-skills. The skill is part of the prototyping-testing plugin. The Discover tab under /plugin offers the same choice, and Gemini CLI users can copy the repository's extensions folder. The prototyping-testing plugin also ships /prototyping-testing:evaluate.
What are Nielsen's ten usability heuristics in this skill?
Visibility of system status, match between system and real world, user control and freedom, consistency and standards, error prevention, recognition over recall, flexibility and efficiency, aesthetic and minimalist design, error recovery, and help and documentation. The agent checks the interface against each one.
What is the severity scale in heuristic-evaluation?
A 0 to 4 scale. Zero is not a usability problem, 1 is cosmetic only, 2 is a minor problem, 3 is a major problem that is important to fix, and 4 is a catastrophe that must be fixed before release. Each documented issue carries one of these ratings.
How many evaluators does heuristic-evaluation recommend?
Three to five. The best practices say multiple evaluators find more issues and that each should evaluate independently before comparing. They also advise focusing on real user tasks, suggesting solutions as well as problems, and combining the review with real user testing.
What license does heuristic-evaluation use?
It is MIT licensed, as is the whole Owl-Listener/designer-skills repository. The pack holds 273 skills and 76 commands across 33 plugins, and heuristic-evaluation is one of the skills in the prototyping-testing plugin. The nine design practice plugins, prototyping-testing among them, ship from this repository.
When should I use heuristic-evaluation instead of design-critique?
Use heuristic-evaluation when you need findings without recruiting participants. Its description points to design-critique in design-ops for a facilitated team feedback session. The evaluation produces rated issues from an expert pass, and the critique gathers feedback from a group. The skill's best practices also say to combine the review with real user testing.