Evaluate user-facing interfaces against Nielsen's 10 Usability Heuristics. Complements /mycelium:service-check (Downe = service-level quality, Nielsen = interface-level quality).
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Evaluate interface-level usability using Jakob Nielsen's 10 heuristics (1994; titles unchanged, definitions refined 2020). This complements Downe's 15 Good Services principles: Downe checks end-to-end SERVICE quality, Nielsen checks screen-level INTERFACE quality. Both are needed for user-facing work.
The system should always keep users informed about what is going on through timely and appropriate feedback.
The system should speak the users' language, using familiar words, phrases, and concepts rather than system-oriented terms.
Users often perform actions by mistake. They need a clearly marked "emergency exit" to leave the unwanted action.
Users should not have to wonder whether different words, situations, or actions mean the same thing.
Even better than good error messages is a careful design that prevents problems in the first place.
Minimize the user's memory load by making elements, actions, and options visible.
Accelerators — unseen by the novice user — may speed up the interaction for the expert user.
Dialogues should not contain information that is irrelevant or rarely needed.
Error messages should be expressed in plain language, precisely indicate the problem, and constructively suggest a solution.
Even though it is better if the system can be used without documentation, it may be necessary to provide help.
Nielsen's heuristics are diagnostic principles. They do not cover the quantitative interaction-design layer: Fitts's/Hick's target-and-choice math, Gestalt grouping (proximity, similarity, closure, common-region), typography (line-length/CPL, line-height, F-pattern scanning), and cognitive-load laws. If this project has a UI surface and the user wants that layer, the haabe/ux-axioms-mcp server covers it. Per its README, the server exposes ~44 such axioms with analyze_ui / suggest_axioms / generate_spec / generate_tests tools — confirm the current install/run path against the repo before recommending a specific setup command.
Mycelium nudges, never installs: offer it, let the user wire it up (per ${CLAUDE_PLUGIN_ROOT}/harness/design-principles.md — autonomy/JIT-tooling). Don't double-surface — where a Nielsen heuristic and an axiom overlap (e.g. heuristic 4 Consistency ↔ Similarity; heuristic 8 Minimalist ↔ Occam's Razor), cite once. The persuasion axioms in that registry are dual-use; the same ethics line as anti-pattern #10 (Dark Pattern Marketing) applies.
## Usability Heuristic Evaluation
| Heuristic | Status | Issues |
|-----------|--------|--------|
| 1. Visibility of status | pass/partial/fail | ... |
| 2. Match real world | pass/partial/fail | ... |
| 3. User control & freedom | pass/partial/fail | ... |
| 4. Consistency & standards | pass/partial/fail | ... |
| 5. Error prevention | pass/partial/fail | ... |
| 6. Recognition > recall | pass/partial/fail | ... |
| 7. Flexibility & efficiency | pass/partial/fail | ... |
| 8. Aesthetic & minimal | pass/partial/fail | ... |
| 9. Error recovery | pass/partial/fail | ... |
| 10. Help & documentation | pass/partial/fail | ... |
Overall: [X]/10 passing
Priority fixes: [list most impactful issues]
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Accessibility audit, scoped to the surfaces a product actually has. Detects web / rendered_markdown / terminal / native_app / video_audio / document / headless, then applies only the criteria that bind. WCAG 2.1 AA in full for web; not at all for headless.
日本語の概要は準備中です。原文の説明を表示しています。
Bring Mycelium into a project that already has code. Detects that the repo predates the framework, asks before touching anything, then reads the codebase to draft what it CAN establish (delivery, solution shape) and — the actual point — names what it cannot (purpose, strategy, real user evidence). The output is a discovery backlog with a head start, never a filled canvas.
日本語の概要は準備中です。原文の説明を表示しています。
Design the smallest viable test to validate or invalidate a critical assumption. Based on Torres's assumption testing framework, organized by Gilad's AFTER model (Assessment → Fact-Finding → Tests → Experiments → Release Results).
日本語の概要は準備中です。原文の説明を表示しています。
Use before any research activity or significant decision. Reviews cognitive biases relevant to the current stage.
日本語の概要は準備中です。原文の説明を表示しています。
Use to evaluate whether current work aligns with Better Value Sooner Safer Happier. Run at diamond completion and periodically.
日本語の概要は準備中です。原文の説明を表示しています。
Lint canvas files for staleness, missing fields, inconsistent evidence types, and orphaned references. Run periodically or before major transitions.
日本語の概要は準備中です。原文の説明を表示しています。