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 southleft/figma-console-mcp-skills --skill figma-extract-design-systemgit clone --depth 1 https://github.com/southleft/figma-console-mcp-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/southleft/figma-console-mcp-skills/figma-extract-design-system)<a href="https://agentmods.dev/skills/southleft/figma-console-mcp-skills/figma-extract-design-system"><img src="https://agentmods.dev/badge/skills/southleft/figma-console-mcp-skills/figma-extract-design-system/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/southleft/figma-console-mcp-skills/figma-extract-design-system"><img src="https://agentmods.dev/badge/skills/southleft/figma-console-mcp-skills/figma-extract-design-system.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 78 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00133 | $0.02250 |
| Opus 5 | $0.00067 | $0.01125 |
| Sonnet 5 | $0.00027 | $0.00450 |
| Haiku 4.5 | $0.00013 | $0.00225 |
Grade A, and why
figma-extract-design-system 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 12d 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 — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
figma-extract-design-system — production app(s) → design system
Organizations let small teams vibe-code products with AI. The products work — but there's no design system underneath, so nothing is consistent, reusable, or versionable. This skill runs the extraction engagement: app(s) → component inventory → DTCG tokens → Storybook package → (optionally) Figma.
Two valid stopping points:
- Code-led org: stop at the Storybook package. It becomes the versioned source of truth that gets packaged and re-imported into the products.
- Design-led org: continue into Figma — tokens become variables (
figma_import_tokens), key components become Figma component sets.
Tooling modes
- Preferred — figma-console-mcp connected (Local Mode): use the five
figma_ds_*tools. They are deterministic, bounded, and persist manifests under<outDir>/.extraction/so the engagement survives session boundaries. - Fallback — no console MCP: do the same phases yourself with Read/Grep/Bash (slower, less
deterministic). Keep the same artifacts: an
analysis.jsonyou write yourself, DTCGtokens/tokens.json, and the same package layout. For the Figma round-trip use thefigma-import-tokensskill (official Figma MCPuse_figma).
Phase 0 — Discovery (talk first)
Establish with the user before scanning:
- Target app path(s) — absolute paths. Multiple apps is a feature: the real design system is the intersection of their products.
- Code-led or design-led? Decides whether Phase 6 (Figma) happens.
- Brand/naming context — product name, existing brand color names if any. Needed for the semantic-naming pass.
- Where the design system lives — default
<first app>/design-system, or a separate dir in the monorepo.
Phase 1 — Analyze
figma_ds_analyze { targets: ["/abs/path/app"], outDir: "/abs/path/app/design-system" }
Review the summary with the user before proceeding — especially the
architecture block, which is what separates a design system from a UI kit:
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.
- 12d ago First seen · 159 lines · 133 tokens per session scan A c18f10671aab
figma-extract-design-system is a skill published in the GitHub repository southleft/figma-console-mcp-skills (81 stars, last pushed 26d ago), licensed MIT. It adds 133 tokens to every session and 2,250 once invoked, about $0.0007 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 2026-08-30.
Other skills, from other repositories
design-to-code
Use this skill when translating UI/UX designs into production-ready frontend code. Handles design source ingestion (Figma, Sketch, Adobe XD, screenshots, design specs), design token extraction, component hierarchy mapping, responsive breakpoint strategy, accessibility-first implementation (WCAG 2.1 AA minimum)…
design-to-code-check
Check alignment between a specific design specification and its code implementation — a focused, single-component or single-screen comparison. Trigger when someone says: does this match the design, check implementation, design code alignment, what's different between the design and the build, spec check…
design-tokens
Sets up, validates, resolves, and exports design tokens per the DTCG (Design Tokens Community Group) Format Module 2025.10 standard, and owns all colour contrast measurement and remediation — every text/background pair on both APCA Lc and the WCAG ratio via tokens contrast, with a fix that moves OKLCH lightness only.…
better-design
Build, improve, and review production interfaces with the Better Design MCP. Use for frontend design, UI, UX, interactions, flows, forms, navigation, components, styling, design systems, accessibility, responsive layouts, motion, React Native, Expo, Shopify, Three.js, and product video work, especially when creating…
prototype
Make an idea lookable and clickable at the cheapest fidelity that answers its one open question — an ASCII sketch, an SVG click-dummy with hotspots linking screens, or a self-contained HTML file. Use when the user wants to see or try an idea before committing to structure, tokens, or copy. Triggers on prototype…
layout-rules
A battle-tested avoid-list of 39 structural and interaction defect classes for tool/dashboard UIs (structure, footers/metrics, tables, color/contrast, URL state, JS traps, accessibility/touch, i18n/theming, de-slop), distilled from a real build/audit cycle. Use when designing, building, or reviewing any layout…