Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/MichelKerkmeester/skilled-agent-harness_spec-driven-loopsnpx agentmods add skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/sk-design-md-generatorWrote 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/michelkerkmeester/skilled-agent-harness_spec-driven-loops/sk-design-md-generator)<a href="https://agentmods.dev/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/sk-design-md-generator"><img src="https://agentmods.dev/badge/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/sk-design-md-generator/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/michelkerkmeester/skilled-agent-harness_spec-driven-loops/sk-design-md-generator"><img src="https://agentmods.dev/badge/skills/michelkerkmeester/skilled-agent-harness_spec-driven-loops/sk-design-md-generator.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 | $0.00030 | $0.09053 |
| Opus 5 | $0.00015 | $0.04526 |
| Sonnet 5 | $0.00006 | $0.01811 |
| Haiku 4.5 | $0.00003 | $0.00905 |
Grade C, and why
sk-design-md-generator scanned grade C with 1 finding 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 yesterday.
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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- Keywords: design system, design tokens, css extraction, design.md, website design extraction, design reference, tokens.json, playwright, design-to-markdown, design-system generator, css tokens, color extraction, typ How it starts
The opening of the file, as written. The whole thing — 421 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Design System Extractor (md-generator)
Captures a live website's real, measured CSS into a publication-quality DESIGN.md — a v3 Style Reference: a named, role-driven, ship-ready design-system handoff (named colour tokens, semantic type scale, named components, Surfaces, Elevation, Agent Prompt Guide, Similar Brands, and copy-paste Quick Start CSS + Tailwind) that AI agents build against without hallucinating colors, fonts, spacing, or shadows. Runs a three-phase pipeline (extract, write, validate) through an embedded Playwright crawler that samples five viewports and emits verbatim tokens.json. Deep operational detail lives in references/.
Boundary. This skill is the extraction and format-fidelity engine for design references. It captures what already exists — it does not decide values for a surface that does not exist yet, which belongs to the sibling skill
sk-design. The transport —mcp-figma— moves design data; this skill produces the authoritative measured reference that transport and downstream implementers (sk-code) consume. Where both skills apply to the same surface, a measured reference from here outrankssk-design's defaults; where no measurement exists,sk-designdecides.
1. WHEN TO USE
Activation Triggers
Use when the user wants to:
- Extract a live website's real CSS into a machine-readable
DESIGN.mdv3 Style Reference, with every hex, pixel, font-weight, radius, and shadow copied verbatim from the running page. - Produce a hallucination-proof design document for an AI agent to code against —
DESIGN.mdreplaces guesses with measured ground truth. - Capture colors (including dark-mode palette if detected), typography stacks, shadows, border-radii, spacing scales, CSS custom properties, framework markers, icon-system hints, motion tokens, and a11y data across five responsive viewports.
- Validate a generated
DESIGN.mdagainst its sourcetokens.jsonto confirm hex accuracy and section completeness. - Generate visual validation reports (HTML preview, diff report) from a
DESIGN.md+tokens.jsonpair. - Re-extract a site after a redesign to update the design-system reference.
What ships with it
60 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.
- assets/cardinal-rules-card.md 3.3 KB
- assets/design-md-prompt-template.md 5.0 KB
- assets/source-of-truth-router-card.md 4.1 KB
- backend/.npmignore 136 B
- backend/package-lock.json 78 KB
- backend/package.json 1.0 KB
- backend/README.md 9.4 KB
- backend/scripts/a11y-extract.ts 14 KB runs code
- backend/scripts/build-write-prompt.ts 8.0 KB runs code
- backend/scripts/cli.ts 3.2 KB runs code
- backend/scripts/cluster.ts 60 KB runs code
- backend/scripts/corpus-baseline-v3.ts 12 KB runs code
- backend/scripts/crawl.ts 34 KB runs code
- backend/scripts/css-analyzer.ts 19 KB runs code
- backend/scripts/css-tree.d.ts 1.8 KB runs code
- backend/scripts/dark-mode-detect.ts 9.4 KB runs code
- backend/scripts/design-boundary-detect.ts 12 KB runs code
- backend/scripts/dom-collector.ts 21 KB runs code
- backend/scripts/extract.ts 28 KB runs code
- backend/scripts/formatters-v3.ts 16 KB runs code
- backend/scripts/framework-detect.ts 7.3 KB runs code
- backend/scripts/guided-run.ts 14 KB runs code
- backend/scripts/icon-detect.ts 9.4 KB runs code
- backend/scripts/interaction-capture.ts 32 KB runs code
- backend/scripts/motion-extract.ts 6.1 KB runs code
- backend/scripts/output-policy.ts 3.6 KB runs code
- backend/scripts/preview-gen.ts 12 KB runs code
- backend/scripts/proof.ts 20 KB runs code
- backend/scripts/README.md 12 KB
- backend/scripts/render-safety.ts 4.3 KB runs code
- backend/scripts/report-gen.ts 38 KB runs code
- backend/scripts/schema-v3.ts 25 KB runs code
- backend/scripts/study-exemplars.ts 24 KB runs code
- backend/scripts/study-prepare.ts 7.5 KB runs code
- backend/scripts/types.ts 15 KB runs code
- backend/scripts/typography-role-v3.ts 3.2 KB runs code
- backend/scripts/validate.ts 34 KB runs code
- backend/tests/a11y-extract.test.ts 1.1 KB runs code
- backend/tests/build-write-prompt.test.ts 6.0 KB runs code
- backend/tests/cluster-classify.test.ts 2.0 KB runs code
- backend/tests/cluster.test.ts 13 KB runs code
- backend/tests/corpus-baseline-v3.test.ts 3.1 KB runs code
- backend/tests/css-analyzer-transitions.test.ts 2.2 KB runs code
- backend/tests/design-boundary-detect.test.ts 3.6 KB runs code
- backend/tests/fixtures/README.md 1.7 KB
- backend/tests/fixtures/study-cases.ts 5.8 KB runs code
- backend/tests/formatters-v3.test.ts 5.1 KB runs code
- backend/tests/guided-run.test.ts 5.4 KB runs code
- backend/tests/icon-detect.test.ts 2.6 KB runs code
- backend/tests/motion-extract.test.ts 3.6 KB runs code
- backend/tests/output-policy.test.ts 3.5 KB runs code
- backend/tests/parseargs.test.ts 1.4 KB runs code
- backend/tests/README.md 5.9 KB
- backend/tests/render-safety.test.ts 3.6 KB runs code
- backend/tests/report-preview-overwrite.test.ts 5.4 KB runs code
- backend/tests/schema-v3.test.ts 11 KB runs code
- backend/tests/study-exemplars.test.ts 14 KB runs code
- backend/tests/typography-role-v3.test.ts 1.9 KB runs code
- backend/tests/validate.test.ts 14 KB runs code
- backend/tsconfig.build.json 267 B
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.
- yesterday Changed 0a4289c9b7f3
- 8d ago First seen · 421 lines · 30 tokens per session scan C ee3c1874aa32
sk-design-md-generator is a skill published in the GitHub repository MichelKerkmeester/skilled-agent-harness_spec-driven-loops (35 stars, last pushed yesterday), licensed MIT. It adds 30 tokens to every session and 9,053 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (hidden instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
chakra-ui-builder
Build responsive, accessible UI components and layouts using Chakra UI v3, install or configure Chakra UI in new and existing projects, and design scalable themes using tokens, semantic tokens, recipes, and slot recipes. Use this skill whenever a user asks to build, create, or generate any UI component, page, form…
visual-ralph
Visual Ralph orchestration for frontend UI from generated references, static references, or live URL targets, using $ultragoal with built-in visual verdict and pixel-diff evidence until the implementation matches and leaves a reproducible design system.
frontend-ui-dark-ts
Build dark-themed React applications using Tailwind CSS with custom theming, glassmorphism effects, and Framer Motion animations. Use when creating dashboards, admin panels, or data-rich interfaces with a refined dark aesthetic.
reveal-hover-effect
Build cursor-following spotlight reveals that expose a second aligned image through a soft radial mask. Use for hover-to-color, before-and-after, x-ray, material, texture, product-detail, and illustrated hero effects where a desaturated or embossed base image should remain visible while another treatment follows an…
frontend-visual-qa
Audits already-rendered web, landing-page, HTML deck/slide, browser tool/game, dashboard/admin, design-system, and desktop UIs using real-browser or native-app journeys, inspected screenshots, DOM geometry, responsive or projection viewports, and a bundled Playwright sweep. Use after UI implementation to find…
prototype-web
A clickable, high-fidelity web product prototype with navigation, a hero section, feature cards, steps, social proof, and optional pricing. It is designed to resemble a finished landing page while remaining a prototype.