# Design Evaluation Audit > Scores designs on cognitive checklists with ranked fixes. - Skill: `lyndonkl/design-evaluation-audit-2` (Agent Skill, multi-file: 4 files) - Install (CLI): `npx skillmds@latest add lyndonkl/design-evaluation-audit-2` - Raw SKILL.md: https://api.skillmd.com/api/skills/lyndonkl/design-evaluation-audit-2/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Security - License: MIT - Author: lyndonkl (https://skillmd.com/u/lyndonkl) - Updated: 2026-09-21 - Page: https://skillmd.com/skills/lyndonkl/design-evaluation-audit-2 --- # Design Evaluation & Audit Evaluates an existing design against cognitive science principles with a repeatable 8-dimension checklist (visibility, hierarchy, chunking, simplicity, memory, feedback, consistency, scanning), a 4-criteria visualization audit scored 1-5 (clarity, efficiency, integrity, aesthetics), and severity-classified fix recommendations. The output is an audit report with evidence, not a redesign; `cognitive-design` handles creating designs and `cognitive-fallacies-guard` handles misleading charts. ## When to Use - Conducting design reviews or critiques: "review this design", "evaluate", "audit", "what's wrong with this". - Evaluating a design for cognitive alignment against repeatable checklists rather than taste. - Quality assurance before launch, or diagnosing usability issues in a shipped interface, form, or dashboard. - Choosing between design alternatives with objective, scored criteria. - Not for creating new designs (`cognitive-design`) or detecting misleading visualizations (`cognitive-fallacies-guard`). Load any reference file with `skill_view("design-evaluation-audit", file_path="references/.md")`; the paths below are relative to this skill. --- ## Skill Boundaries **Use instead of this skill for:** - Creating new designs from scratch: use `cognitive-design` - Learning cognitive theory: use `cognitive-design` Path 1 - Detecting misleading visualizations: use `cognitive-fallacies-guard` --- ## Design Review Workflow **Time:** 30-90 minutes depending on scope **Copy this checklist and track your progress:** ``` Design Evaluation Progress: - [ ] Step 1: Systematic Assessment - [ ] Step 2: Visualization Quality Audit (if applicable) - [ ] Step 3: Severity Classification & Prioritization - [ ] Step 4: Fix Recommendations ``` ### Step 1: Systematic Assessment Apply the Cognitive Design Checklist across all 8 dimensions: Visibility, Visual Hierarchy, Chunking, Simplicity, Memory Support, Feedback, Consistency, Scanning Patterns. Check every item. Record pass/fail for each dimension with specific evidence. **Resource:** [Cognitive Design Checklist](references/cognitive-checklist.md) ### Step 2: Visualization Quality Audit (if applicable) If the design includes data visualizations, apply the 4-Criteria Visualization Audit. Score each criterion 1-5: Clarity, Efficiency, Integrity, Aesthetics. Calculate average and identify weakest dimension. **Resource:** [Visualization Audit Framework](references/visualization-audit.md) ### Step 3: Severity Classification & Prioritization Classify every finding by severity: - **CRITICAL:** Integrity violations, accessibility failures, users cannot complete core tasks. Fix immediately. - **HIGH:** Clarity/efficiency issues preventing use, missing feedback for critical actions, working memory overload (>10 ungrouped items). Fix before launch. - **MEDIUM:** Suboptimal patterns, aesthetic issues, minor inconsistencies. Fix in next iteration. - **LOW:** Minor optimizations, polish items. Fix when convenient. **Priority rule:** Fix foundation-first — perception before coherence, integrity before aesthetics, critical before high. ### Step 4: Fix Recommendations For each finding, document: 1. **What is wrong** — specific description with evidence 2. **Why it matters** — which cognitive principle is violated 3. **How to fix** — concrete, actionable recommendation 4. **Expected outcome** — what improves after the fix 5. **Effort estimate** — quick fix (minutes), moderate (hours), significant (days) Verify fixes don't harm other dimensions. --- ## Path Selection Menu ### Path 1: Run Cognitive Design Checklist **Choose this when:** Evaluating any interface, layout, content page, form, or general design. **What you'll get:** Pass/fail across 8 cognitive dimensions, test methods, common failures, severity-classified findings. **Time:** 20-40 minutes **→ [Go to Cognitive Design Checklist](references/cognitive-checklist.md)** --- ### Path 2: Run Visualization Audit **Choose this when:** Evaluating data visualizations — charts, graphs, dashboards, infographics. **What you'll get:** 1-5 scores on Clarity, Efficiency, Integrity, Aesthetics with pass/fail threshold. **Time:** 15-30 minutes per visualization **→ [Go to Visualization Audit Framework](references/visualization-audit.md)** --- ### Path 3: Combined Review **Choose this when:** Comprehensive review covering both interface elements and data visualizations. **Process:** Run Cognitive Checklist first, then Visualization Audit on each data component, merge findings, produce unified fix list. **Time:** 45-90 minutes **→ Start with [Cognitive Checklist](references/cognitive-checklist.md), then [Visualization Audit](references/visualization-audit.md)** --- ## Quick Reference ### 3-Question Rapid Check **1. Attention** — "Is it obvious what to look at first?" - If NO: hierarchy and visibility issues **2. Memory** — "Is the user required to remember anything that could be shown?" - If NO: memory support and chunking issues **3. Clarity** — "Can someone unfamiliar understand in 5 seconds?" - If NO: simplicity and comprehension issues All YES = likely cognitively sound. Any NO = run full checklist on the failing area. --- ## Verification An audit report is complete when: - All 8 checklist dimensions have a pass/fail with specific evidence, and every visualization has four 1-5 scores (see [Scoring Summary](references/cognitive-checklist.md#scoring-summary) and [Using the 4-Criteria Framework](references/visualization-audit.md#using-the-4-criteria-framework)). - Every finding carries all five fields: what is wrong, why it matters (the violated principle), how to fix, expected outcome, effort. - Findings are ordered CRITICAL, HIGH, MEDIUM, LOW, foundation-first. - Scoring the report against [assets/evaluators/rubric_design_evaluation.json](assets/evaluators/rubric_design_evaluation.json) averages 3.5 or higher, with Severity Classification and Evidence Quality both at 3 or above. Write the report to a file with `write_file` and run `readability-check` on it before delivery. --- ## Guardrails **Out of scope:** Creating designs, teaching theory, providing domain guidance, replacing user testing, or covering full accessibility compliance. **In scope:** Systematic evaluation against cognitive principles, severity-classified findings, prioritized fix recommendations, and visualization quality scoring.