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 memi-design/design-skills --skill token-architecturegit clone --depth 1 https://github.com/memi-design/design-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/memi-design/design-skills/token-architecture)<a href="https://agentmods.dev/skills/memi-design/design-skills/token-architecture"><img src="https://agentmods.dev/badge/skills/memi-design/design-skills/token-architecture/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/memi-design/design-skills/token-architecture"><img src="https://agentmods.dev/badge/skills/memi-design/design-skills/token-architecture.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 | $0.00039 | $0.00736 |
| Opus 5.5 | $0.00016 | $0.00294 |
| Sonnet 5.5 | $0.00008 | $0.00147 |
| Haiku 4.5 | $0.00004 | $0.00074 |
Grade A, and why
token-architecture 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 29d 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 — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Token architecture
When to use
Use when a design value affects more than one component, theme, or platform, or when Figma/Paper values disagree with code tokens. For a single screen implementation, begin with design-to-code-handoff; this skill handles the token decision within that workflow. Do not introduce a new tier system merely because the current naming differs from an example.
Establish the source of truth
- Read the repository's design-system docs, token files, generated outputs, theme code, and any Figma variable mapping. Identify the actual source of truth and its owner; do not assume Figma or code always wins. Distinguish authored files from generated files and do not edit generated output directly.
- Inventory consumers: CSS custom properties, Tailwind configuration, component styles, Storybook stories, native outputs, and design variables where present. Record which are in scope and which cannot be inspected.
- Capture the exact design source and revision or timestamp, relevant mode, token name, value, and usage. A screenshot color sample is an approximation, not a verified token reference.
Decide the smallest change
| Design use and source | Existing semantic token | Consumers and modes | Decision | Owner and unresolved difference |
|---|
- Reuse an existing semantic token when the intent and supported modes match, even if a raw value differs slightly in one screenshot. Explain material visual differences.
- Extend an existing token or add a semantic alias when multiple consumers share a new intent. Keep primitives and component-specific values in the layers the repository already uses.
- Add a component token only when that component needs a distinct stable contract. Avoid one token per screenshot value and avoid bypassing the system with hardcoded values.
- Defer when ownership, modes, or the intended semantic meaning is unknown. Record the decision needed rather than inventing a color or propagating an unreviewed design change.
What ships with it
1 file 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.
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.
- 29d ago Changed · -280 lines · +7 tokens per session 9318f5f05155
- 1mo ago First seen · 321 lines · 32 tokens per session scan A 640f471421d0
token-architecture is a skill published in the GitHub repository memi-design/design-skills (7 stars, last pushed 1mo ago), licensed MIT. It adds 39 tokens to every session and 736 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-03.
Other skills, from other repositories
design-to-code-check
Compares a component or screen's design spec with its code and logs each gap as a build error or spec gap. Use it for design QA, handover review or sign-off: whenever someone asks if a build matches its design or spec, even one component with the spec pasted in. System-wide drift: drift-detection.
design-engineering
Premium design engineering skill for agentic workflows — produces high-end, distinctive UI designs using DESIGN.md as the portable contract across Pencil MCP (in-IDE canvas), Figma MCP (team handoff + design tokens), and Google Stitch (vibe exploration + AI generation). Enforces anti-generic principles, WCAG 2.2 AA…
building-components
The workflow for building, rebuilding, or restyling any design-system component (Button, Dialog, Select, Table, Tabs, Tooltip, and so on). Trigger whenever the user says "let's do the dialog", "restyle X", "bring in / pull in a component", "build the select", "rebuild this atom", or starts work on an item from the…
gsp-color
Design color systems — palettes, contrast, semantic mapping, dark mode — use when: pick colors, build a palette, semantic tokens, dark mode pairings, brand color, accent color.
gsp-typography
Design type systems — scale, pairing, fluid type, vertical rhythm — use when: pick fonts, font pairing, type scale, vertical rhythm, fluid typography, body and heading sizes.
analyze-component-set-figma
Analyze a Figma COMPONENTSET as a state machine for code generation — extract variant axes (state/size/etc.), map state variants to CSS pseudo-classes (hover→:hover, focus→:focus-visible, disabled→:disabled, error→[aria-invalid]), and compute per-variant visual diffs (only what changes per state). Use when generating…