Home›People›UX

UX · evaluate against 20 heuristics + emotional arc

UX

UX review agent · evaluates student-facing screens against 20 heuristics (Nielsen 10 adapted + 8 education/AI-specific) · H-1 first-time-student simulation

Uplifted · M-2 L-3 · M-3 done
  • Agent file.claude/agents/ux.md @ cbd7bbfc
  • Expertise corpusdocs/agent-knowledge/ux/expertise/ @ 053c171f
  • Team membershipUplift cohort · Specialized direct-instance roster
  • Direct-instance worktreeaiuni-uplift-ux on s75-uplift-ux
  • ToolsRead · Grep · Glob
  • Catalog sectionAGENT-ROLES-AND-RESPONSIBILITIES-CATALOG §3.9
  • Existence statusexists

Summary

UX review agent. Evaluates student-facing screens against 20 heuristics (Nielsen's 10 adapted + 8 education/AI-specific). Catches usability issues that PO and QA miss — wrong labels, inappropriate placeholders, mobile breakage, emotional regression, accessibility violations, cognitive load problems.

Runs alongside QA Phase 3 — both must pass. Non-negotiable rule 1: H-1 first-time-student simulation runs FIRST on every screen, before any structural / heuristic check. Non-negotiable rule 2: H0 content-readability word-by-word runs SECOND. Template-variable / internal-jargon / developer-syntax leak in rendered text = automatic CRITICAL. WCAG 2.2 AA hard gate — contrast ratios with measured hex values; font sizes in px; touch targets in px; ARIA + focus + keyboard validated. Cross-references Designer corpus `wcag-2.2-accessibility.md` for the comprehensive treatment.

Distinguishing characteristicsDifferent from Designer because: UX evaluates against existing patterns, Designer originates new direction. Different from QA because: UX evaluates human experience, QA verifies code/runtime. Different from Validator because: UX scope is product surfaces, Validator scope is curriculum content.

Research portfolio10 primary-literature · 4 clusters

Engagement floorCluster A (heuristic foundation) is non-optional on any visual review; Cluster C (AI-specific) is required when AI is in the loop (most AI Uni surfaces); Cluster B (education-specific) applies on curriculum surfaces; Cluster D (emotional/affective) applies on session-flow / emotional-arc evaluation per `docs/agent-knowledge/ux/expertise/named-authors-checklist.md`.

Cluster A · Heuristic-evaluation foundation

Domain: Nielsen ten heuristics · severity scale · ISO international standard · scanning reader

  • Nielsen, J. (1994). Usability Engineering. Morgan Kaufmann. + Nielsen, J., & Molich, R. (1990). Heuristic evaluation of user interfaces. CHI '90.
    10 usability heuristics · 4-point severity scale · scanning reader · independent-evaluator floor
    nielsen-heuristic-evaluation.md @ 053c171f
  • ISO (2020). ISO 9241-110:2020 — Ergonomics of human-system interaction — Part 110: Interaction principles. International Organization for Standardization.
    7 dialogue principles · international-standard cross-validation of Nielsen heuristics · 2020 revision
    iso-9241-110-2020.md @ 053c171f

Cluster B · Education-specific UX

Domain: pedagogical dimensions · learnability-usability synergy · engagement-as-construct

  • Reeves, T. C., & Reeves, P. M. (1997). Effective dimensions of interactive learning on the World Wide Web. In Khan, B. H. (Ed.), Web-based instruction (pp. 59-66). Educational Technology Publications.
    10 pedagogical dimensions for learning-system evaluation · pedagogical philosophy · learning theory · goal orientation · task orientation
    reeves-reeves-1997-educational-dimensions.md @ 053c171f
  • Squires, D., & Preece, J. (1999). Predicting quality in educational software: Evaluating for learning, usability and the synergy between them. Interacting with Computers, 11(5), 467-483.
    Synergy framework · learnability-usability dual-evaluation · pedagogical-mismatch findings
    squires-preece-1999-learnability-usability-synergy.md @ 053c171f
  • Quinn, C. N. (2005). Engaging Learning: Designing e-Learning Simulation Games. Pfeiffer.
    8 engagement dimensions · engagement-as-evaluable-construct · learner-driven simulation
    quinn-2005-engaging-learning.md @ 053c171f

Cluster C · AI-specific interaction

Domain: AI-UX guidelines · explainability question-bank · production-pattern catalog

  • Amershi, S., Weld, D., Vorvoreanu, M., Fourney, A., Nushi, B., Collisson, P., et al. (2019). Guidelines for Human-AI Interaction. Proceedings of CHI 2019, ACM.
    18 AI-UX guidelines across 4 phases (initially · during interaction · when wrong · over time) · validated empirically
    amershi-2019-ai-design-guidelines.md @ 053c171f
  • Liao, Q. V., Gruen, D., & Miller, S. (2020). Questioning the AI: Informing Design Practices for Explainable AI User Experiences. Proceedings of CHI 2020, ACM.
    XAI Question Bank · 9 question categories · how/why/what-if explainability framework
    liao-gruen-miller-2020-explainability.md @ 053c171f
  • Microsoft (2023). HAX Toolkit / AI Patterns Catalog. Microsoft Research.
    Production-pattern catalog · AI mistake handling · feedback loops · user trust calibration
    microsoft-2023-ai-patterns.md @ 053c171f

Cluster D · Emotional/affective layer

Domain: three levels of emotional design · flow theory

  • Norman, D. A. (2004). Emotional Design: Why We Love (or Hate) Everyday Things. Basic Books.
    Three levels — visceral / behavioral / reflective · emotional design distinct from usable design
    norman-2004-emotional-design.md @ 053c171f
  • Csikszentmihalyi, M. (1990). Flow: The Psychology of Optimal Experience. Harper & Row.
    Flow channel · challenge-skill balance · 8 conditions for flow · anxiety/boredom anti-zones
    csikszentmihalyi-1990-flow.md @ 053c171f

Cross-class references

  • Cluster · Norman 2013 seven-stages-of-action
    → designer corpus norman-affordance-and-action-cycles.md @ 6d571734
    Cross-class to Designer corpus Cluster 1 — UX agent applies seven-stages diagnostic as auxiliary to H-1 first-time-student simulation. NOT duplicated in UX corpus per Q4 disposition (L-2 precedent).
  • Cluster · Sweller CLT + Mayer multimedia learning
    → researcher corpus @ 9e61df0c
    UX agent applies cognitive-load lens on instructional surfaces. Cross-references Researcher + Designer corpora rather than duplicating.
  • Cluster · WCAG 2.2 AA
    → designer corpus wcag-2.2-accessibility.md @ 6d571734
    UX agent applies WCAG as standing gate; cross-reference Designer corpus `wcag-2.2-accessibility.md` for the comprehensive treatment. UX corpus references the standard via lens-rubric Q4 + anti-patterns-catalog #4.

Significant project contributionsS60+ baseline → S75 P0-1

  • S75 P0-1 · 2026-05-05 · L-3 cascade lane
    Authored canonical UX expertise corpus at `aiuni-uplift-ux` commit `053c171f` — 10 primary-literature files spanning 4 clusters + named-authors-checklist + 13-entry anti-patterns catalog (incl. #13 cross-class decorative-invocation per L-2 Q4 precedent) + 7-question lens-rubric. PO+Code-PM Tier 2a R29 ceremony PASS clean. — 053c171f
  • S75 P0-1 · 2026-05-05 · M-3 codification
    UX agent file rewrite at commit `cbd7bbfc` — references Canonical Expertise Corpus (lines 17-28) + 11-step engagement protocol + Q1-Q7 lens-rubric self-application before substantive review. — cbd7bbfc
  • S75 P0-1 · 2026-05-05 · cascade-resume Stage 1 dual-lead
    Co-authored Stage 1 mockup deliverable for AI Uni Corporate Intranet (this surface) as Designer + UX dual-lead per HANDOFF v4 §6 Step 1 amendment. Single-terminal dual-persona mode. Output: 7 HTML mockups + design-tokens.md + IA-flow.md + lens-rubric self-application doc. — 4a821231
  • S75 (R30 codification) · 2026-05-05 · C15 sign-off-block discipline
    R30 NEW: UX review surface-back targeting user-authorization gate MUST emit formal sign-off block (reviewer + timestamp + scope reviewed + issues + verdict (SIGN OFF / BLOCK / SIGN OFF WITH NOTES) + handoff-back disposition). Phrase-level PASS = anti-pattern #11.
  • S60+ baseline · 2026-04-15 · ux/principles/ baseline
    5 existing AI-Uni-context principle files preserved per Q1 disposition (L-1 precedent) — principles are AI-Uni-applications of canonical primary literature documented in expertise/ corpus.

Learningsauthored + cross-cutting

Authored

  • H-1 first-time-student simulation runs FIRST on every screen
    Before any structural / heuristic check. If section is not obvious from on-screen alone, section FAILS regardless of other heuristics. Krug 2014 first-law floor + Norman 2013 seven-stages diagnostic as auxiliary.
  • Anti-pattern #5 — confirmation-bias-on-author's-design
    Evaluator sees prior agent's reasoning before evaluating; produces findings biased toward author's frame. Sibling C24 lead-with-user-summary.
  • Anti-pattern #11 — soft-sign-off-after-R30-codification
    Phrase-level PASS verdict instead of formal sign-off-block (reviewer + timestamp + scope + verdict). R30 codified after S75 C15 fired.
  • Anti-pattern #13 — decorative cross-class invocation
    Citing Researcher / Designer corpus literature decoratively without primary-literature anchor in cross-class claim = explicit failure. Cross-class refs require commit-pin + structural verification.

Cross-cutting applied

  • SHARED-LEARNINGS PRINCIPLE 16 — internal identifiers never render to end users
    UX H0 anti-pattern #8 — template-variable / internal-jargon / developer-syntax leak in rendered text = automatic CRITICAL.
  • R30 — formal sign-off block discipline
    Surface-back targeting user-authorization gate MUST emit formal sign-off block. S75 C15 origin.
  • R32 + R34 + R29 Step 0 amendment
    Anchor-grounding granularity + research-input-vs-design-target distinction + target-correctness statement. Cross-cutting failure-mode triangulation closes C19+C24+C25.

Skills + hooksused + constraining

Skills used (always-loaded)

Primary · 20-heuristic evaluation

ux-review

20-heuristic UX evaluation · WCAG checks · emotional arc · cognitive load assessment · H-1 simulation.

AI Uni · brand

design-system

AI Uni visual identity, brand tokens, component patterns. Consumed when reviewing visual-token decisions.

Hooks constraining

  • Tool-scope · agent file frontmatter
    Read · Grep · Glob ONLY (no Write at agent-file level). UX is review-only, never modifies files.
  • Lens-Rubric self-application (M-3 codification)
    Q1-Q7 honest self-application before any UX surface-back. Phrase-level PASS = anti-pattern #11; primary-literature anchor required per question.
  • R30 sign-off-block discipline (S75 C15 codification)
    Formal sign-off block on review surface-backs targeting user-authorization gate (reviewer + timestamp + scope + issues + verdict + handoff-back disposition).

Last updated · refresh details

  • Profile auto-regenerated2026-05-06T12:50:00Z
  • Refresh cadenceDaily midnight UTC · per-session-end · on-demand button (admin-tier optional)
  • agent file.claude/agents/ux.md @ cbd7bbfc
  • corpusdocs/agent-knowledge/ux/expertise/ @ 053c171f
  • catalogdocs/agent-knowledge/AGENT-ROLES-AND-RESPONSIBILITIES-CATALOG.md @ d7d1aee8+241affb8

Refresh strategy per HANDOFF v4 §12 — profile auto-regenerates from agent-file + corpus + executions + git log.

Cross-references from the relationship graph

Edges sourced from cross-refs/matrix.json · also browsable on the org chart.

Cross-references 2 skills · 2 hooks

Skills · always-loaded

  • Ux-Review
    20-heuristic Nielsen-adapted framework + H-1 first-time-student simulation.
  • Design-System
    Brand tokens consumed when reviewing visual-token decisions.

Hooks constraining

  • tool-scope-frontmatter · hard
    Read · Grep · Glob ONLY
  • lens-rubric-self-application · process-discipline
    Q1-Q7 before any review surface-back