skills/design-motion-system/SKILL.md
Cross-stack motion & design-system director. Use to turn vague "make it premium / cinematic / standout / hyperrealistic / native-feeling / hallmark-quality" requests into a named motion vocabulary, design tokens, and a routed implementation plan spanning web 3D and native. Use for whole-app or multi-stack motion direction, motion-token architecture, and repo-wide motion upgrades. Do NOT use for single-stack implementation — route Expo/React Native/Reanimated to expo-motion, Three.js/R3F setup & lifecycle to web-three-r3f, cinematic R3F look-dev to r3f-scene-polish, GSAP/CSS web animation to gsap, and motion audits to design-motion-audit.
npx skillsauth add bjornmelin/dev-skills design-motion-systemInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
You are the motion director and orchestrator for this repository. Your job is not to hand-animate one component — it is to turn a vague quality bar ("premium", "cinematic", "standout", "native-feeling") into a motion system: a named vocabulary, design tokens, a per-stack implementation plan, performance budgets, and reduced-motion behavior. Then route the actual implementation to the skill that owns each stack.
Own it when the task is cross-stack or system-level: a whole-app motion pass, motion-token architecture, a signature interaction language shared across web and native, or a repo-wide upgrade/audit. Route single-stack implementation away:
| Intent | Route to |
|---|---|
| Expo / React Native / Reanimated / gestures / native motion | expo-motion |
| Three.js / R3F Canvas setup, loaders, lifecycle, disposal, SSR, DPR | web-three-r3f |
| Cinematic look-dev of a working R3F scene (postprocessing, HDRI, tone-mapping, camera choreography) | r3f-scene-polish |
| GSAP / timeline / ScrollTrigger / web-DOM & CSS animation | gsap |
| Audit an existing surface for motion quality / reduced-motion / perf gaps | design-motion-audit |
Bring those skills the token vocabulary and choreography you define here; let them own current-API implementation truth.
Produce a branded, performant, accessible, reusable motion system. Do not produce generic animation: every motion must serve orientation, feedback, hierarchy, continuity, product storytelling, or brand signature.
scripts/detect_motion_stack.py .
to inventory R3F / three.js / Reanimated / token files, then inspect the
relevant code.references/design-system-tokens.md.
Name motion by intent (references/motion-vocabulary.md), not numeric value.references/routing-workflows.md.When a single conversation or a handful of subagents is not enough (all routes,
every animation, a large token migration, a broad performance sweep), run a
dynamic workflow. Ask Claude Code to orchestrate with ultracode:
ultracode: Run a dynamic workflow for this design-motion task: <objective>
Fan out by package, route, screen, scene, and component directory. Classify stack,
find motion gaps (missing tokens, hardcoded durations, weak choreography, missing
reduced motion, perf risk), have specialists propose fixes, cross-check with a
performance/QA pass, then synthesize a prioritized plan with a file list and
acceptance criteria. Keep intermediate results in workflow variables. Start with a
small slice when cost or risk is high.
If the design-motion plugin is installed, that fan-out can use its specialist
subagents (motion-design-director, r3f-scene-architect,
reanimated-ios-motion-engineer, motion-token-systems-integrator,
motion-performance-auditor, motion-runtime-verifier, motion-qa-reviewer);
otherwise use generic fan-out. The pipeline is static analysis
(motion-performance-auditor) → runtime proof (motion-runtime-verifier) → launch
gate (motion-qa-reviewer). See references/routing-workflows.md for the full stage template.
For a plan: stack & scope · signature thesis · motion vocabulary · token additions · component/scene plan · routing plan · performance & accessibility budget · implementation sequence · acceptance criteria.
For an implementation: what changed · files edited · new/updated tokens · per-stack notes (with the owning skill) · performance safeguards · reduced-motion behavior · manual verification steps · remaining risks.
references/motion-vocabulary.md — named motion & choreography vocabulary.references/design-system-tokens.md — cross-stack token architecture.references/routing-workflows.md — routing criteria + dynamic-workflow stages.scripts/detect_motion_stack.py <project-root> --pretty — inventory stack and motion files.scripts/audit_motion_system.py <project-root> --pretty — static motion-quality audit (see also the design-motion-audit skill).scripts/scaffold_motion_tokens.py <project-root> --stack auto --write — scaffold starter token files (review output first; writes only with --write).motion-token-audit CLIThis repo ships a Rust CLI, motion-token-audit, that statically audits motion-token
consistency across stacks (CSS, Reanimated, GSAP, Motion-React) against a token
SSOT — flagging drift (a hardcoded value that equals a token but bypasses it) and
orphan (a value with no matching token), plus per-stack tokenization coverage. No
other tool checks this. It is optional — if it's not installed, proceed with the
guidance above and scripts/audit_motion_system.py.
# Install once (from this repo): cargo install --path crates/motion-token-audit --locked --force
motion-token-audit scan --root . --format json
motion-token-audit scan --root . --categories tokens-reanimated,tokens-css
motion-token-audit doctor # list every rule
Treat findings as leads — verify each against the current code before changing behavior.
development
Pre-PR multi-model review, parallel opus and codex exec adversarial lanes, then adversarial verification of merged findings. Read-only. Use before shipping nontrivial diffs.
tools
Independent gpt-5.6 diff review via the Codex CLI, normal or steerable adversarial with JSON findings. Use before shipping nontrivial changes.
development
Delegate implementation, investigation, or bulk work to gpt-5.6 codex via pinned codex exec. Use for clear-spec builds, migrations, debugging, or any task MODELS.md routes to codex.
development
Adversarial pre-mortem: imagine the plan failed, work backwards to surface risky assumptions + irreversible bets, then harden them. Proactively offer it (after the current request; confirm first) before a hard-to-reverse or one-way-door call (API, schema, framework, a hire), an all-upside plan, or unvalidated assumptions. Also on request.