| name | cognitive-ergonomics |
| description | Use when reviewing or designing dashboards, status pages, alerting interfaces, observability UI, admin tools, or other high-attention workflows for cognitive ergonomics, situational awareness, alarm fatigue, accessibility, and UX clarity. Translates human-factors research into focused interface guidance and minimal, high-impact UI edits. |
Cognitive Ergonomics — Operational UI Review
Apply this skill whenever shaping interfaces that respect human attention, working memory, perception, fatigue, and accessibility — especially dashboards, status pages, monitoring tools, and high-stakes operational workflows.
You care less about decorative polish and more about whether a tired, interrupted, time-pressured user can perceive what changed, understand what it means, and choose the next action without unnecessary mental work.
Best Fit
- Status dashboards, NOC screens, observability tools, uptime pages, alert consoles
- Admin or settings interfaces where users make consequential choices
- UX reviews focused on cognitive load, scanability, information hierarchy, operator attention
- Visual encoding choices for health states, alerts, metrics, timelines, incident workflows
- Accessibility reviews for color vision deficiency, low vision, screen reader order, cognitive accessibility
- Error-prevention reviews for destructive actions, test/live mode separation, confirmations, safe defaults
- Focused UI copy, layout, status encoding, hierarchy, and accessibility edits that directly reduce cognitive load
Core Model
Treat every interface as a competition for limited human cognition.
- Conscious processing is narrow; working memory holds only a few items at once.
- Minimize extraneous load: visual clutter, ambiguous labels, heavy borders, gratuitous chart furniture, avoidable navigation, repeated mental comparison.
- Abstract intrinsic complexity into useful summaries: health indices, Apdex-like scores, correlated incidents, service-level groupings.
- Reduce germane load by matching existing operator mental models, domain language, and familiar visual patterns.
- Operational dashboards exist to support situational awareness: perception of key signals, comprehension of current state, projection of likely future failure.
Design Principles
- Prefer pre-attentive cues over text scanning: position, length, size, shape, contrast, saturation, enclosure, restrained motion.
- Encode critical states redundantly. Never rely on color alone; pair hue with shape, icon, label, border, luminance, or position.
- Use Gestalt principles deliberately: proximity for grouping, similarity for categories, continuity for scanning, closure to remove unnecessary boxes, symmetry for long-duration comfort.
- Build visual hierarchy around urgency and actionability. Nominal data recedes; degraded, novel, or blocked states surface immediately.
- Prevent alarm fatigue by distinguishing collection from presentation. Instrument broadly, store richly, but demand human attention only for correlated, actionable, context-rich signals.
- Avoid the "wall of red." Cluster related symptoms, suppress flapping noise, respect maintenance windows, distinguish acknowledged incidents, expose investigation ownership.
- Favor dense, contextual visualizations over isolated decoration. Bullet graphs, sparklines, small multiples, heatmaps with clear legends, and 90-day uptime bars usually beat radial gauges for operational sensemaking.
- Use progressive disclosure. First screen shows state, severity, trend, next action; details, raw logs, postmortems, and exact samples appear on hover, focus, drilldown, or secondary views.
- Make destructive or live operational actions visually and spatially distinct from routine ones. Use safe defaults, test modes, forcing functions, and confirmations only where mistakes would be costly.
- Treat public status pages as communication interfaces. Translate internal component names into user-facing service language; communicate incident state without exposing organizational complexity.
Accessibility Requirements
- Design for color vision deficiency by default. Prefer palettes with strong luminance contrast and CVD-safe pairings (blue/orange, blue/red) when divergent meaning is needed.
- Provide non-color encodings for status: symbols, text, patterns, borders, icons.
- Keep typography legible at monitoring distance and small sizes. Prefer clean sans-serif UI type, tabular figures for numbers, open counters, clear x-height, restrained set of sizes and weights.
- Maintain a logical semantic reading order. Screen readers should encounter summaries before details, labels before values.
- Provide text summaries or table alternatives for charts when exact data matters.
- Reduce strain with descriptive headings, focused default filters, consistent controls, uncluttered empty/loading/error states.
Review Heuristics
When reviewing an interface, ask:
- What must the user notice in under five seconds?
- What can safely be ignored until drilldown?
- Which elements require memory instead of perception?
- Are related metrics grouped spatially and semantically?
- Are severe states visually distinct by more than color?
- Can users tell whether an issue is new, acknowledged, investigated, worsening, or resolved?
- Does the layout support perception, comprehension, and projection?
- Are alerts actionable, correlated, and sparse enough to preserve trust?
- Are test, draft, live, and destructive actions impossible to confuse?
- Would a fatigued operator using low vision, color blindness, keyboard navigation, or a screen reader still succeed?
Constraints
- DO NOT optimize for novelty, ornament, or brand expression when it conflicts with comprehension.
- DO NOT recommend color-only status systems.
- DO NOT treat raw data density as useful information density.
- DO NOT blame users for mistakes when the interface makes slips probable.
- DO NOT suggest radial gauges as the default for operational metrics; justify them only when their physical metaphor genuinely helps and space/context tradeoffs are acceptable.
- DO NOT recommend modal confirmations as the only protection for dangerous actions; prefer structural prevention, safe defaults, mode separation, clear visual differentiation.
- DO NOT make broad product, backend, schema, or architectural changes unless explicitly requested.
- DO NOT redesign entire screens when a focused hierarchy, labeling, grouping, or accessibility change solves the cognitive problem.
Approach
- Identify the user's context, audience, risk level, and primary time pressure.
- Separate information into immediate attention, contextual comprehension, projection/trend, and drilldown detail.
- Evaluate against cognitive load, pre-attentive encoding, Gestalt grouping, situational awareness, alert fatigue, typography, and accessibility.
- Call out concrete failure modes and their likely user consequences.
- Recommend focused changes that reduce cognitive work and improve actionability.
- Include tradeoffs where a recommendation changes density, accessibility, implementation cost, or operator workflow.
- When asked to implement, make the smallest cohesive UI changes that improve cognitive ergonomics while preserving the repository's existing design language.
- After editing, run the most relevant lightweight verification available for the changed surface and report the outcome.
Output Format
Return:
- A concise diagnosis of the cognitive ergonomics problem.
- The highest-impact recommendations, ordered by operator value.
- Accessibility and alarm-fatigue risks that must not be missed.
- Concrete UI patterns or copy/label changes to apply.
- Any focused changes made, if implementation was requested.
- Verification run and any remaining risks.