Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add zakariaf/Flutter-Skills --skill design-system-structuregit clone --depth 1 https://github.com/zakariaf/Flutter-SkillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/zakariaf/flutter-skills/design-system-structure)<a href="https://agentmods.dev/skills/zakariaf/flutter-skills/design-system-structure"><img src="https://agentmods.dev/badge/skills/zakariaf/flutter-skills/design-system-structure/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/zakariaf/flutter-skills/design-system-structure"><img src="https://agentmods.dev/badge/skills/zakariaf/flutter-skills/design-system-structure.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00211 | $0.04462 |
| Opus 5 | $0.00105 | $0.02231 |
| Sonnet 5 | $0.00042 | $0.00892 |
| Haiku 4.5 | $0.00021 | $0.00446 |
Grade A, and why
design-system-structure scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 195 lines — stays where its author put it; the contents beside it link to each section on GitHub.
design-system-structure
A design system is code that is incapable of holding a stray opinion: every aesthetic value lives in one directory, and every widget reads it back through a named slot. This skill owns how tokens, themes, components, and shapes are structured and consumed — not which colors or radii to pick (that belongs to your design source of truth). Getting the structure right makes a reskin a diff lib/theme/ and turns a stray Color(0xFF…) in feature code into a build failure instead of a style nit.
Read the reference for the task at hand:
references/token-tiers-and-themeextension.md— two-tier token naming, ThemeExtension mechanics, the assertingof(), hand-authoredColorSchemevsfromSeed, honestlerp.references/motion-and-reduced-motion.md— the three animations Material mounts by default, whyNoSplashis not enough, reduced-motion-means-zero,pumpAndSettlebans in tests.references/typography-and-fonts.md— bundling fonts,LicenseRegistry,FontWeightdriveswght, silentFontVariationno-ops, subsetting without instancing, per-script fallback cascades.references/contrast-and-redundant-encoding.md— AA as a unit test, ≥3 grayscale-legible signals per meaning, reading a11y flags fromMediaQuery.
Run scripts/check_raw_values.sh and scripts/check_font_bundling.sh before a PR.
Non-negotiable rules
- No raw aesthetic value outside
lib/theme/**— the gate is the law. AColor(0x…)/Colors.*(excepttransparent) /Curves.*/Duration(milliseconds:|seconds:(exceptDuration.zero) / literalBorderRadius.circular(n)/fontSize: nin feature or shared UI code failsscripts/check_raw_values.sh. A legitimate new need is a new token slot, never a// ignore— one place to diff is the whole point. - Two token tiers; widgets read only the semantic tier. Tier 1 = primitives named by the measured value (
neutral12= OKLCH lightness ×100), never by rank (grey700), appearance (darkGrey), or brand (brandPrimary). Tier 2 = semantic slots named by role (surface,onSurface,hairline,accent). A widget that reaches a Tier-1 primitive has hardcoded one theme. WHY: rank scales have no room to insert and lie in dark mode; appearance names invert catastrophically; brand names die with the brand. - Read tokens through a
ThemeExtensionand an assertingof(context). Never a staticconstclass, never anInheritedWidgetside channel.of()asserts on a missing extension — it does not?? fallback. WHY:MaterialAppdrives light/dark switching and correct-by-construction invalidation; a fallback silently ships a theme no test verified, and loud-in-debug beats wrong-in-field. - Hand-author
ColorScheme; neverColorScheme.fromSeed(ordynamic_color) for identity.fromSeed's per-role overrides do not propagate — overridesurfaceandsurfaceContainerHighstays seed-derived. State every role you consume, keep the M3 role names so Material widgets theme themselves, and let unstated roles takeColorScheme's own defaults, not a seed's opinion. WHY: "seed plus a few overrides" means owning the ~9 roles you chose and every seam of the ~40 you didn't. - Layer the system: tokens → theme → modifiers/components → shapes. Author light and dark by hand (dark is never an auto-flip), each passing AA independently, and attach every extension to both
ThemeDatas. Inject the theme exactly once, at the composition root. WHY: theming the app must touch one directory; a palette-only dark theme ships light's ratios into the dark. lerphonestly or snap deliberately;copyWithis mandatory. Interpolate colors and doubles; snap non-interpolables (font families,ShapeBorder) att < 0.5. If your app bans theme animation and you snap everything, leave a comment — a bare step-lerpreads as unfinished and the next reader will "fix" it. WHY: a new field silently forgotten inlerpis the classic design-system rot.- Color is a derived token, computed LAST — never a status's only channel. Color is a pure function of a canonical value object (an enum/small class), resolved through a semantic slot at the end; status is never read from color, and the supporting color is a slot read, not a literal. The never-color-alone floor itself — ≥3 non-color signals (glyph, label, weight/position) per stateful meaning — is owned by
accessibility-as-code; this skill only enforces that the color reinforcing them is a derived slot. WHY: a color not derived from the value object drifts out of sync, and a widget that leans on color alone dies in grayscale. - Reduced motion collapses to zero, not "gentler." Read
MediaQuery.disableAnimationsOf(context)and resolve toDuration.zero/ a cross-fade to the end state — never a shorter duration or a softer curve. Motion is never the only signal for a state change. WHY: a user who asked the OS to stop animations asked for stop. - Restore the persisted theme before first paint. Load the saved theme mode / variant before
runApp, from the same versioned settings the rest of the app uses (not an asyncshared_preferencesread that lands a frame late). An unknown or corrupt stored value falls back explicitly and visibly, never to anullthat leavesof()asserting on a device with no debugger. WHY: a flash of the wrong theme is a sudden luminance change the user did not cause. - Bundle fonts; never
google_fontsat runtime. Declare fonts inpubspec.yaml, ship the license text and register it viaLicenseRegistry, and drive weight withFontWeight(which drives thewghtaxis) — not a redundantFontVariation('wght', …). Declare a per-scriptfontFamilyFallbackcascade that ends in a known-good face. WHY:google_fontsships an HTTP code path by default;FontVariation('opsz'|'ital', …)silently no-ops on a font that lacks the axis, so a "working" italic is invisible in review. - Painters snapshot tokens;
paint()never readsBuildContext. ResolveTheme.of(context).extension<T>()once at the widget layer and pass the values into theCustomPainter/Scenefields. WHY: keepspaint()allocation-free and testable without aMaterialApp; seecustom-canvas-and-gestures. - The reduced-motion flag is read from
MediaQuery(disableAnimationsOf), never app state — it feedsresolveMotion(this skill's helper, below). Reading the other platform a11y flags (boldTextOf,highContrastOf,textScaler) fromMediaQueryrather than a stale app-state copy is owned byaccessibility-as-code.
What ships with it
9 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- examples/app_theme.dart 8.3 KB
- examples/reduced_motion.dart 4.0 KB
- examples/status_encoding.dart 2.7 KB
- references/contrast-and-redundant-encoding.md 4.5 KB
- references/motion-and-reduced-motion.md 4.6 KB
- references/token-tiers-and-themeextension.md 6.5 KB
- references/typography-and-fonts.md 6.5 KB
- scripts/check_font_bundling.sh 1.1 KB runs code
- scripts/check_raw_values.sh 1.8 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 195 lines · 211 tokens per session scan A cae59a80a8d3
design-system-structure is a skill published in the GitHub repository zakariaf/Flutter-Skills (2 stars, last pushed 10d ago), licensed MIT. It adds 211 tokens to every session and 4,462 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
accessibility
Audits or remediates Flutter widgets against WCAG 2.2 conformance levels A, AA, or AAA across iOS, Android, Web, macOS, Windows, and Linux, covering Semantics labels and screen reader output under VoiceOver and TalkBack, touch target sizes, dragging alternatives, focus order and keyboard navigation, color contrast…
android-edge-to-edge
Migrate a Jetpack Compose app to edge-to-edge display and fix system bar inset issues. Use when UI components are obscured by navigation/status bars, fixing IME insets, or enabling edge-to-edge for SDK 35+.
common-mobile-animation
Apply motion design principles for mobile apps covering timing curves, transitions, gestures, and performance-conscious animations. Use when implementing screen transitions, gesture-driven interactions, shared-element animations, or optimizing animation frame rates on iOS, Android, or Flutter.
common-mobile-ux-core
Enforce universal mobile UX principles for touch-first interfaces including touch targets, safe areas, and mobile-specific interaction patterns. Use when building mobile screens, handling touch interactions, or validating safe area compliance.
flutter-design-system
Enforce Design Language System adherence in Flutter. Use when implementing design tokens, preventing hardcoded colors/spacing, or building a DLS.
material-theming
Best practices for Flutter theming with Material 3, treating ThemeData as the single source of truth for colors, typography, component styles, and spacing. Use when creating, modifying, or reviewing ThemeData, ColorScheme, TextTheme, component themes, spacing systems, or light/dark mode support, and whenever widget…