orbyn-ui-design ยท diff
git:20261031.d5a1c25 to git:20261031.2390887
2 added, 0 removed. Audit A to A.
---
name: orbyn-ui-design
description: "Design, implement, or review Orbyn UI and UX across web, desktop, and mobile: task flows, concise copy, responsive layouts, overlays, and AI interactions. Use for visible product flows; not backend-only work."
---
# Orbyn UI and UX Design
Apply Orbyn Engineering's UI approach to the target product: make its task flow easy
to scan and operate at real window sizes. Use the agreed palette and
shared components. Compact default text must remain readable and grow with user
accessibility settings.
For social announcement graphics, product screenshot compositions and feed
mockups, use [social-preview](../social-preview/SKILL.md), with
[social-content](../social-content/SKILL.md) for the post. This skill owns changes
to the product UI itself; a promotional mockup is not an implemented screen.
For colour-direction references, semantic light/dark roles or contrast review,
use [reference palettes and semantic colour](references/colour-palettes.md).
Keep an existing project brand first; optional references are not automatic theme
adoption. Its gallery separates raw labels, derived roles and contextual checks.
For code-driven UI motion and an editable preview, use the existing
[motion composition and inspector](examples/spatial-review/motion/README.md).
Keep live gesture velocity, interrupted presence, keyboard takeover and reduced
motion separate from deterministic export seeks. Optional graphics/audio backends
never become prerequisites for ordinary product UI.
## Read before designing
- Read `AGENTS.md` and the agreed task scope; preserve scope and stage order.
- For product copy and information hierarchy, read
`skills/orbyn-content-design/SKILL.md`; keep detailed content rules there.
- Read [design rules](references/design-rules.md) for typography, layout, copy,
AI behavior, and the visual acceptance matrix.
- Read [UX rules](references/ux-rules.md) when changing navigation, task steps,
forms, onboarding, feedback, recovery, or explanatory content.
- Read [research](references/research.md) when choosing or revising a rule. It
distinguishes accessibility requirements, research guidance, and Orbyn choices.
- Inspect the owning screen and shared controls before adding a local override.
Resolve token/component owners from the consuming project's stack guide and
manifests. In Orbyn's application these are `packages/core/src/theme/presentation.ts`,
`desktop/src/styles/global.css`, and `mobile/src/theme/index.ts`; other projects
use their own owners. Never create these paths to fit the skill.
## Resolve product direction
Read the consuming checkout's design/brand pointers in its optional
[project context](../orbyn-setup/references/project-context.md), then inspect the
actual token and component owners. Preserve its agreed palette, type scale,
density, navigation and motion policy. This library supplies no universal blue
accent, modal layout or pixel scale. When a guide is absent, use the existing
interface and the user's brief; record unsettled choices before a broad redesign.
Read [prototyping guidance](references/prototypes.md) when an interactive sketch
would resolve a design question. Use `orbyn-research` for unknown platform or
accessibility requirements; do not replace existing product tokens with a foreign
host's theme or tool APIs.
For new visual direction, a requested redesign or a design-quality audit, read
[visual direction](references/visual-direction.md): derive composition from the
person's task and real material, choose a coherent system and distinguish defects
from stylistic preferences. For existing UI changes, use the
[redesign contract](references/redesign-contract.md) to preserve routes, content,
forms, permissions and recovery behavior. Audit-only requests produce findings
without product edits. A small polish does not authorize a broad redesign.
For validation with people, use [user evidence](../orbyn-research/references/user-evidence.md)
to choose nonleading representative tasks and accessible participation. Separate
observed usability from demand/adoption or causal benefit; [product discovery](../orbyn-plan/references/product-discovery.md)
owns that outcome decision. For model-powered behavior, use
[AI application engineering](../orbyn-ai-application/SKILL.md) for grounding,
quality and tool contracts; UI keeps capability expectations, source/uncertainty,
correction, dismissal and manual continuation visible. Feedback is not training
consent and does not enable automatic capture.
For substantial workspace composition, collections, editors or review surfaces,
use [workspace composition](references/workspace-composition.md). Decide defaults,
primary labels and competing panels before decoration; preserve saved preferences
and use its compact task-based acceptance receipt in the existing change note.
## Design and implement
1. State the person's main task, starting state, and observable success. Map the
steps, cancellation, failure, and return path before arranging controls.
2. Use existing text roles, spacing, and controls. Reduce competing headings,
repeated actions, and explanatory copy before changing font sizes.
3. Adapt to the actual content width: collapse secondary panels, stack forms,
or use a drawer. Keep every action reachable without squeezing the main task.
4. Preserve data and authority: drafts, selections, permissions, approvals,
consent, and backend run states must survive visual changes.
5. Cover the supported surfaces affected by the requested change, using their
platform-appropriate navigation and controls. Apply parity only where the
product contract requires it; record agreed platform limitations.
## Review and delivery
Follow the consuming project's visual-review policy and the changed surface.
Choose [evidence surface](references/prototypes.md#choose-the-evidence-surface-before-building)
before building a harness or capturing UI; identify the loaded candidate/window
and adapters, honoring actual-app intent or reporting its blocker.
Use proportional evidence, avoiding a blanket sweep for unrelated changes. Use
tests/API evidence for functional behavior and reuse valid prior UI evidence.
For a material layout/interaction risk or reproduced
visual defect, inspect the relevant flow when authorized. If an independent
visual role is requested, use the session/subagent topology in
[coordination](../orbyn-workflow/references/project-policy.md#delegation-and-host-configuration). Give that inspector the candidate checkout/ref,
URLs, exact states, viewport sizes, themes, and a Markdown report destination.
Specify whether its assignment is capture-only or includes analysis, following
the user's current authorization. Report delegated pixel review as delegated;
do not label it personal inspection.
Choose only cases relevant to the named risk; do not expand into a full matrix
of unchanged states or screens. Stop when those cases are checked. Verify actual CSS viewport,
zoom, and screenshot provenance before diagnosing a cropped or scaled image.
Passing builds, DOM dimensions, and screenshot metadata alone do not prove a
usable layout. Review density, discoverability, and task completion as well as
containment. Browser review does not prove installed-native behavior.
Fix reproduced problems and request a targeted recheck. Retain failed attempts
and unverified states. If browser permission or export is rejected, report the
blocked action and use an allowed handoff; do not bypass it through another
browser, port, transport, or indirect export. Integrate a scoped checkpoint only
when its acceptance evidence supports it; do not claim the whole task complete.
## Worked example
[Recovering a failed save](examples/failed-save.md) applies a consuming project's
style and draft/conflict contract without claiming an unperformed visual check.
+ For assistant conversation/composer surfaces, use [assistant-chat UI](references/assistant-chat.md) to connect the existing runtime, typed parts, scoped drafts, reading intent and product-owned tokens. Ordinary forms or comments do not require a chat runtime.
+
## Async flow ownership
For fetch/save/search or subscription changes, use [asynchronous state and recovery](references/async-state.md). Preserve request identity, stale-result handling, drafts, current scope and cleanup in the actual framework.
## Experience rule
Evaluate the complete task from the person's starting state through success,
cancellation and recovery. Resolve trade-offs around that outcome, preserving
accessibility and authority. Use the [experience method](references/principles/experience-first.md)
when a local visual improvement could make the end-to-end task harder.
## Deliver a preview the human can use
When presenting new or changed UI for human design review, offer or reuse the
live feedback method in [review packages](references/review-packages.md), without
requiring a separate invocation. Reuse an existing component preview for a small
button change; respect audit-only requests and avoid rebuilding unchanged pages.
For requested interactive comments, settings or durable records, deliver that path.
Cover each affected page,
component/instance and control; bind feedback to actual source/state and separate
browser draft, submitted feedback, repo save and accepted design decisions. Keep
requested durable source/settings/reviews with the consuming design owner, then
reconcile its existing design document and focused regressions within scope.
For a requested prototype or preview, produce the actual app route/story or a
self-contained HTML artifact as appropriate, open it with supported host tools
when possible, and provide the verified link plus states to try. Follow the
[prototype method](references/prototypes.md) for deterministic demo controls,
simulation labels, rendering evidence and transfer into the real frontend/backend.
For stateful flows, use [behavior preview](references/behavior-preview.md) to
connect relevant scenarios, state/rule IDs and actual checks. Hand lifecycle
statecharts to `architect` and critical property checks or drift to `tdd`; retain
visual, model and implementation evidence separately. Small visual work stays small.
Use the [artifact lifecycle](../orbyn-workflow/references/artifact-lifecycle.md)
for storage, server lifetime and cleanup. A source file or screenshot alone does
not establish that the human can exercise the interaction.
For positive page identity, fresh structure after navigation, or supplied Figma
node/assets-to-primitives mapping, use [design intake](references/design-intake.md).
For bidirectional, translated or live-locale interfaces, use
[localization and RTL](references/localization-rtl.md), preserving actual language,
direction, focus/input and native-surface evidence separately.
## Motion and mobile surfaces
For ordinary transitions/gestures use [motion](references/motion.md) and measured [motion performance](references/animation-performance.md). For browser touch/keyboard/viewport use [mobile web](references/mobile-web.md), and real schema/data extremes use [stress fixtures](references/stress-fixtures.md). Demonstrated Expo/RN screens select [native UI](references/mobile/native-ui.md), [navigation](references/mobile/navigation.md), [data/lifecycle](references/mobile/data-and-lifecycle.md) or [UI-runtime motion](references/mobile/motion.md) only as relevant.