feat: add seven UX-laws lenses (complete the /ux-laws-review roster)

[vigneshsubbiah16]

Summary

• Adds the seven remaining UX-laws lenses to complete the /ux-laws-review roster. Combined with trust-and-honesty (shipped in PR feat: add /ux-laws-review parent + trust-and-honesty lens #2), every law in the 30 Laws of UX (re-weighted for 2026 AI-native development) now lives in exactly one canonical lens.
• All 10 S-tier laws are accounted for with zero double-coverage; the parent skill's surface-aware preset routing resolves cleanly for all 5 surface presets.
• Each lens file mirrors the canonical trust-and-honesty.md template; identical JSON output contract across all 8 lenses so the parent skill assembles a unified scorecard at /build Step 4.2 without per-lens schema awareness.

The seven lenses

Lens	Laws (⭐ = S-tier)	Evidence vocabulary	Routes to
speed-and-flow	Doherty ⭐, Flow ⭐, Goal-Gradient, Zeigarnik, Parkinson	Quantitative — INP/LCP p75, waterfalls, progress affordances	checkout, AI chat, full
cognition	Cognitive Load ⭐, Working Memory ⭐, Miller, Chunking, Serial Position	Structural — decision counts, chunk-boundary analysis	onboarding, dashboard, settings, chat
decision-architecture	Choice Overload ⭐, Hick, Occam ⭐, Pareto	Structural + behavioral — choice counts, default presence	checkout, settings, full
attention	Selective Attention, Von Restorff, Paradox of the Active User ⭐	Visual + behavioral — heat-map analysis, power-user-vs-first-timer	onboarding
familiarity	Jakob, Mental Model ⭐	Conceptual — convention adherence, comparison-vs-leaders	onboarding, settings
gestalt-visual	Common Region, Proximity, Similarity, Uniform Connectedness, Prägnanz	Visual — annotated screenshots, grouping violations	dashboard
targeting	Fitts	Quantitative + visual — target size in px/tap area, thumb-reach zones	dashboard

S-tier coverage (10/10, no duplicates)

• speed-and-flow → Doherty Threshold, Flow
• cognition → Cognitive Load, Working Memory
• decision-architecture → Choice Overload, Occam's Razor
• attention → Paradox of the Active User
• familiarity → Mental Model
• trust-and-honesty (already merged) → Tesler's Law, Peak-End Rule

Decisions taken (with council weigh-in)

Reviewed by principal-architect → principal-engineer → code-simplifier → elite-pr-reviewer → principal-architect (final sign-off):

1. Cross-lens law overlap: Hybrid — one canonical lens per law, with italicized read-only *Related law:* cross-references elsewhere. Never re-score a law in a non-canonical lens.
2. Lens size asymmetry: Kept targeting as an intentional one-law lens (~91 lines). Parent routing already wires it as a peer.
3. Template fidelity: Six-section skeleton fixed across all 8 lenses; section contents adapt to lens nature (quantitative / visual / conceptual).
4. Gating thresholds: Uniform across lenses (lens avg < 6 = WARN, < 4 = FAIL). S-tier flag mechanism in the parent — not custom thresholds — handles per-law calibration.
5. Output JSON contract: Identical across all 8 lens files. Non-negotiable for parent-skill consumption.

Cross-references (today, both → Tesler's Law)

• decision-architecture.md Choice Overload row → italicized one-liner pointing readers to Tesler's Law in trust-and-honesty (smart defaults are where the irreducible complexity lives architecturally)
• familiarity.md Mental Model row → italicized one-liner pointing readers to Tesler's Law in trust-and-honesty (pretending the agent's complexity is gone breaks the user's mental model fastest)

Surface-preset routing now resolves cleanly

Preset	Lenses run
checkout	trust-and-honesty + decision-architecture + speed-and-flow
onboarding	familiarity + attention + cognition
dashboard	cognition + gestalt-visual + targeting
settings	cognition + decision-architecture + familiarity
AI chat	trust-and-honesty + speed-and-flow + cognition

All eight referenced lens files exist on disk after this PR lands.

Review findings addressed

elite-pr-reviewer pass 1: 0 CRITICAL, 1 WARNING (self-downgraded as the pattern was consistent across all 3 multi-S-tier files), 11 INFOs (10 confirming-clean observations + 1 actionable).

Action taken: renamed snake_case key paradox_active_user → paradox_of_the_active_user in attention.md to match the architect's plan (3 occurrences). All other observations were confirming the implementation matches the plan.

principal-architect post-fix sign-off: READY TO MERGE — 7 files scoped, paradox rename consistent across all 3 usage sites, S-tier coverage 10/10 with no duplication.

Out of scope (separate follow-up PRs)

• /eng-blocker-council and /ux-blocker-council skills — referenced as forward-references in commands/build.md Step 5 (shipped in PR feat: fan out three reviewers in parallel at /build Step 4 #3); now genuinely the next thing to build
• Modifications to the parent commands/ux-laws-review.md — already stable, governs lens routing

Test plan

Markdown skill library, no automated test framework. Verification by inspection (architect's 7-check plan):

• markdownlint baseline holds: per-file warning count for the 7 new files matches trust-and-honesty.md baseline (same conventions, same long-line / bare-fence style)
• install.sh --force installs all 7 new lenses as namespaced commands (/ux-laws-review:speed-and-flow, etc.)
• Structural consistency: all 8 lens files have exactly 8 H2 headings each
• JSON output contract: identical top-level keys and per-law sub-keys across all 8 files
• Surface-preset routing: all 5 presets in the parent's table resolve to lens files that exist on disk
• S-tier flag accounting: 10/10 S-tier laws covered, each in exactly one lens, no double-coverage
• Single-PR diff sanity: exactly 7 new files under commands/ux-laws-review/, nothing else changed
• Manual smoke: invoke /ux-laws-review against a real PR after merge to confirm fan-out behavior across all eight lenses

🤖 Generated with Claude Code