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 agentmods add skills/genkovich/sdd/design-systemnpx skills add genkovich/sdd --skill design-systemgit clone --depth 1 https://github.com/genkovich/sddWrote 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/genkovich/sdd/design-system)<a href="https://agentmods.dev/skills/genkovich/sdd/design-system"><img src="https://agentmods.dev/badge/skills/genkovich/sdd/design-system.svg" alt="Measured on agentmods" 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 | $0.00213 | $0.02352 |
| Opus 5 | $0.00106 | $0.01176 |
| Sonnet 5 | $0.00043 | $0.00470 |
| Haiku 4.5 | $0.00021 | $0.00235 |
Grade A, and why
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 5d 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 — 132 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: design-system
The design-side twin of survey: once per repo it fixes the design canon in
docs/design-system.md — which tool screens are drawn with (figma / pencil / code), the
platform posture, where tokens come from, and the component inventory new screens must compose.
architecture-map.md §Frontend stays the inventory of the code; this file is the design
canon the design pipeline (ux-flows → screens) and implement/review read. The tool choice
is committed here — never in .claude/sdd.local.md (that file is per-developer and gitignored;
the canon is team-wide).
Question phrasing → ../_shared/ask-style.md. Canon prose follows
artifact_language — frontmatter keys/values (tool:, figma_file:, …), component names and
file anchors stay English → ../_shared/artifact-language.md.
Owner
Whoever owns the product's look (designer / frontend lead / the solo maintainer).
Inputs
- (Optional)
docs/architecture-map.md§Frontend / UI foundation + thefrontend:machine key — the code-side inventory this canon cites. - (Optional) an existing
docs/design-system.md— updated, never silently overwritten. - The design-tool MCPs available in this session (Figma / Pencil), if any → detected in step 3.
Protocol
- Check existing. If
docs/design-system.mdexists, show itstool+ posture + inventory summary and ask: reuse as-is (STOP — nothing to do) / update (continue, edits land in place) / rebuild. Never overwrite silently. - Scan the code side. Read
architecture-map.md§Frontend / UI foundation + thefrontend:key. Map absent, stale, orfrontend: ""on a repo that visibly has UI code → dispatch theexploreragent —subagent_type: "sdd:explorer"(fallbacksubagent_type: "Explore", per../_shared/agent-roster.md) — for the component library / tokens / styling approach / shared primitives, each with afile:line. A repo with no UI code at all is fine — the canon can start tool-side or empty (greenfield: the inventory grows asimplementregisters components). - Detect tools + ask ONE question set. Detect which design-tool MCPs the session actually has
(Figma / Pencil — probe the available tools, don't assume). Then one
AskUserQuestioncall (up to 3 questions in it, phrased per../_shared/ask-style.md): (a) tool — options from what's detected, the detected one first as «(Recommended)»;code(markdown wireframes inscreens.md) is always offered — it needs no MCP and never blocks; (b) platform posture — mobile-first / desktop-first / responsive-both; (c) token source — the code file(s) found in step 2 / the tool's variables / «none yet». - Write the canon. Fill
./templates/design-system.md→docs/design-system.md: frontmattertool+figma_file/pen_file(the one matching the tool; the other stays""), posture, token source, the component inventory — each row citingfile:line(code) or the node/URL (tool-side) — and the cross-screen conventions. - Bootstrap the tool library (
tool: pencilonly — figma links an existing file,codeneeds nothing). The canon naming apen_filethat doesn't exist or is an empty stub is the #1 way this skill «did nothing» — so materialize it: a. File. Absent → write the minimal stub topen_file({"version": "2.17", "children": []}). b. Open + verify. The Pencil MCP ignoresfilePath— it writes into whatever document is active in the app. Ask the user to open the file (macOS:open -a Pen <pen_file>), then verify viaget_app_statethat the active canvas ISpen_file. Mismatch → stop and repeat the ask; never write into a foreign document. c. Seed. Dispatch thepen-keeperagent —subagent_type: "sdd:pen-keeper"(no fallback: without the Pencil MCP this step degrades tocodemode per../_shared/tool-adapters.md) — with the canon path,pen_file, and the token source files. It re-checks the active document, seeds the variables (light + dark), ensures the foundations frame, and returns a reconcile report. d. Persist. Relay the report's user actions — always «Cmd+S in Pencil»: the app holds changes in memory and the MCP cannot save; confirm the file on disk actually changed (size/mtime) before calling the canon done. - Structural self-check — per
../_shared/self-check.md: re-read the file from disk and verify 5 items: (1)tool∈ {figma, pencil, code}; (2) the tool ↔ file fields agree —figma⇒figma_filenon-empty,pencil⇒pen_filenon-empty,code⇒ both""; (3) every inventory row carries a source anchor (file:lineor node/URL); (4) no<placeholder>stubs survive; (5)tool: pencil⇒ thepen_fileexists on disk and is larger than the empty stub (step 5 ran and the user saved). Fix + re-check ≤2 cycles; surface anything unresolved. - Commit + handoff. Propose commit
design-system: establish <tool> canon(on pencil — including the seededpen_file). Then emit the stage-handoff block per../_shared/handoff.md(utility variant) — What I did (incl. «self-check: 4/4 pass») + Review (docs/design-system.md) + Run next: resume your backbone stage (typically/sdd:ux-flows <slug>for the UI feature in flight);/clearoptional.
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.
- 5d ago First seen · 132 lines · 213 tokens per session scan A 0f4c5a5557c3
design-system is a skill published in the GitHub repository genkovich/sdd (119 stars, last pushed today), licensed MIT. It adds 213 tokens to every session and 2,352 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 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…