Agent skill
runi-design-paulbreuler-runi-2
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/runi-design-paulbreuler-runi-2
SKILL.md
runi Design System (Zen & Ryo Lu Principles)
This skill guides creation of UI components and pages following runi's "Zen" design system and Ryo Lu's (Cursor) design principles for high-performance developer tools.
Brand: runi is vigilant, protective, and minimalist. The interface should feel like a "dark room" where content is the only light.
Philosophy: "Collapse uncertainty into truth" through clarity and deliberate focus.
Core Visual Logic (The Tokens)
-
Zen "Ink" Surface Palette:
- App Background:
#080809(Deep Ink) - Surface:
#111113(Standard cards/sections) - Raised:
#1a1a1d(Hover states/elevated UI) - Border Subtle:
oklch(0.2 0.01 240 / 0.4)(Precision 1px strokes) - The "Dark Room" Effect: Use depth through surface shifts rather than shadows.
- App Background:
-
Zen-Signal Mapping (Low-Chroma OKLCH):
- Success (Green):
oklch(0.75 0.1 150)— Verified, spec-compliant. - Warning (Amber):
oklch(0.8 0.1 75)— Drift detected, deprecations. - AI (Purple):
oklch(0.65 0.1 300)— Generated content, suspect. - Error (Red):
oklch(0.65 0.12 25)— Failures, critical debt. - Accent (Blue):
oklch(0.7 0.1 230)— Brand actions (CTAs), focus.
- Success (Green):
-
Zen Inactive State (Window Blur):
- When the window is blurred (
.window-blur), significantly dim all elements. - Borders should drop to very low opacity:
oklch(... / 0.1). - Signals should lose chroma and brightness:
oklch(0.4 0.02 ...). - This maintains the "Dark Room" feel by letting content recede entirely when not in use.
- When the window is blurred (
-
The "Signal" Rule:
- UI should be 90% Grayscale (using the Ink palette).
- 10% Signal: Reserve color strictly for semantic meaning.
- Eliminate "AI slop"—no massive shadows or unnecessary gradients.
-
Typography:
- UI: Inter (system sans-serif) for all interface elements.
- Code/Specs: JetBrains Mono for all data, JSON, and specifications.
- Maintain a clear hierarchy with minimal visual noise.
-
Animations:
- Animations must be "deliberate and calm".
- Use linear easing.
- Durations: 0.2s–0.4s.
Ryo Lu’s Feedback & "Zen" Application
-
Clean Basics (No AI Slop):
- Eliminate unnecessary visual distractions.
- No massive shadows, no decorative purple gradients.
- No "mystery meat" buttons—icons should have clear tooltips or labels.
-
Clarity over Jargon:
- Don't just show technical shorthand (e.g., "DCFs", "400").
- Show the actual problem: "Field 'user_id' is missing from spec" or "Status code differs from OpenAPI definition".
-
State Transparency (Vigilance Monitor):
- Every background state must be visible.
- When AI verification or a background check is running, use a "Vigilance Monitor" progress bar that explicitly explains what is being checked (e.g., "Verifying request against user.yaml...").
-
Progressive Discovery (Avoid Template-itis):
- Reduce initial visual complexity.
- Hide advanced "Building" level tools (AI mapping, semantic search) until the user is deep in a "Room" (a specific spec/request).
- Only show what is necessary for the current context.
Feature-Specific Polish
-
The Request Builder (The Door):
- Lighter than Postman. Focus on raw HTTP method and URL.
- Minimize chrome and UI panels.
-
Drift Sniffer (The Room):
- When a request drifts from the spec, highlight the specific line in the JSON body with a subtle
amber glow (
--color-signal-warning-rgba-15). - Avoid jarring error modals; use in-context visual indicators.
- When a request drifts from the spec, highlight the specific line in the JSON body with a subtle
amber glow (
-
AI Verification (The Building):
- AI-suggested requests/content must be wrapped in a purple border (
--color-signal-ai-rgba-35). - This signifies "Draft" status until explicitly verified or accepted.
- AI-suggested requests/content must be wrapped in a purple border (
Tech Stack
| Component | Technology |
|---|---|
| Runtime | Tauri v2.9.x |
| Backend | Rust 1.80+ |
| Frontend | React 19 + TypeScript 5.9 |
| Editor | CodeMirror (Migrating) |
| Styling | Tailwind CSS 4.x |
| Animation | Motion 12.x |
State Transparency (The Vigilance Monitor)
- Never Surface Errors Silently: Always display Rust/Backend errors in the UI with correlation IDs.
- Visible Background States: If a background process is running (AI verification, spec sync, request execution), it must be visible.
- The Vigilance Monitor:
- Use the
VigilanceMonitorcomponent for long-running or background tasks. - Provide clear, non-jargon labels (e.g., "Verifying response against spec..." instead of "Running DCF check").
- The monitor should use the
vigilance-progressanimation for a calm, flowing feel.
- Use the
Implementation Checklist
- 90% Grayscale (Ink) - Most UI is monochrome deep ink.
- 10% Signal (Zen-Signal) - Color strictly for Amber/Green/Purple/Blue signals.
- Calm Motion - Linear easing, 0.2s-0.4s.
- No Decoration - Eliminate shadows and gradients that aren't functional.
- Contextual Depth - Use surface shifts (
bg-surface->bg-raised) for hover/active states. - Progressive Discovery - Reveal advanced tools only when deep in a specific context.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?