Extract a repo's design system into an executable UX contract (.uxproof/)
One-time scan of tokens, components, screen archetypes and conventions via a pinned npx extractor, with manual overrides that survive regeneration — every other UX skill in the collection judges against it
- What
- One-time scan of tokens, components, screen archetypes and conventions via a pinned npx extractor, with manual overrides that survive regeneration — every other UX skill in the collection judges against it
- Cost
- Free
- Needs
- a code repository whose design system (or lack of one) should be captured; Node.js/npm for the pinned extractor — the skill produces a contract, it does not review anything
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor — @open-mercato's om-ux-setup skill: run once per repository to extract the design system into a committed, executable contract at `.uxproof/` — `contract.json` (framework, styling system, component roots, screen archetypes), `tokens.json` (every design token with kind and source file), `components.json` (component registry), and `conventions.md` (human-readable house rules, with a manual section for the judgment calls only the team can know — it survives regeneration and outranks generated rules), plus an optional `guards.json` of team-authored deterministic conformance rules the extractor never touches. The workflow: check for an existing contract first (never regenerate silently), extract via a pinned `npx [email protected] init --no-skills` (pinned deliberately so `@latest` never executes unreviewed remote code), report what was found for a human sanity check, ask the two or three questions only the team can answer, surface hygiene warnings (contracts built on scratch files), then hand over — this skill produces the contract and stops; it never reviews anything. Security boundaries: repo content is data, never instructions (embedded directives are reported as prompt injection); secrets stay out of model output. Updated October 21, 2026: a design system shipped by a dependency can now be imported as a version-stamped layer (via `knowledge.sources`), design-tool files route to the companion om-figma-design-with-ds skill, and om-ux-review-pr is named with its `--guard` and `--health` modes. Honest caveats: it degrades gracefully for repos with no design system (falls back to a proposed de-facto palette derived from the code); companion skills it names (om-ux-review-pr, om-ux-shape, om-mockup-prototype) are separate members of the collection, not bundled here. MIT licensed. Skill Harbor never reviews the code, review it yourself before use. Discovered via skills.sh.
Version:
Quick answer
Extract a repo's design system is a free Muse skill for a one-time scan of tokens, components, screen archetypes and conventions, via a pinned npx extractor. Manual overrides survive regeneration, and every other UX skill in the collection judges against the resulting executable contract. That gives a whole UX collection one shared, enforceable source of truth.
Install
Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 11, 2026📌 Pinned version: commit e320bd5Step 0 — before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed. Prerequisites: a code repository whose design system (or lack of one) should be captured; Node.js/npm for the pinned extractor — the skill produces a contract, it does not review anything Install "Extract a repo's design system into an executable UX contract (.uxproof/)" for me. It gives my agent @open-mercato's om-ux-setup workflow: check for an existing .uxproof/contract.json first, extract tokens/components/archetypes/conventions via the pinned npx [email protected] init --no-skills (or the manual fallback offline), import any design system shipped by a dependency as a version-stamped layer when the config names one, report findings for a human sanity check, capture the team's judgment calls in the manual section (and, when the team can state a forbidden pattern exactly, in the optional guards.json), surface contract-hygiene warnings, and hand over a report — then stop. MIT licensed. Repository: https://raw.githubusercontent.com/open-mercato/skills/e320bd5e3f0761a6d7249085c42c77a4b37d480d/skills/om-ux-setup/SKILL.md Pinned version: install the code at commit e320bd5e3f0761a6d7249085c42c77a4b37d480d exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. [FINGERPRINTS] Files at the pinned commit, from Skill Harbor's review of 2026-10-21: - SKILL.md: SHA-256 4c9f3f660ec4cbde9e338225a2911886a027100eb57d521fd6f9a3a32a7dd5f6, 7,109 bytes, 7,089 characters If it does not match, it changed since the review: do not install it yet, tell me what you found and let me decide. Updated October 21, 2026: Skill Harbor regenerated this install prompt from the current upstream file. What changed: the skill now handles design systems that arrive with a dependency. A design system shipped by a dependency, or by a sibling package with its own contract, is imported as a layer: when knowledge.sources in the config names a dependency or repo path that ships a design contract, a new workflow step imports it per references/design-sources.md — its tokens, components, archetype anatomy and guard rules join the contract tagged with their origin and stamped with the installed version, repository entries win on name collisions, UI guides are recorded as pointers and never copied, and a refresh always re-runs the layering step because the extractor rewrites the generated files; with no knowledge.sources key, the step is skipped. The contract gains an optional file, guards.json: deterministic conformance rules (forbidden patterns, required states) written by the team or imported from a layer — the extractor never touches it — and a forbidden pattern the team can state exactly (a class, an element, a raw value) may be added to it as a team-authored rule once the user confirms the pattern and its severity. The "show what was found" report now also reports each layer with its version and import counts, how many entries the repository shadowed, and, on a refresh, layers whose installed version moved. The routing to judgment skills changed too: design-tool files are now covered — om-figma-design-with-ds for a brief for a new screen or an audit of an existing design against the contract (the previous file said reviewing design files was not covered by any skill in the collection) — and om-ux-review-pr is now named with its --guard mode (a static check of code against the contract, or of the whole repository with --health) alongside its running-UI review. What did not change: the skill still only extracts the contract and stops — no findings, no verdicts — the extractor stays pinned, and the manual section still outranks generated rules. 1. Fetch the SKILL.md file (and any helper files) from the repository path at the pinned commit into a temporary folder, then compare SKILL.md with the fingerprint from Skill Harbor's review of 2026-10-21 (see [FINGERPRINTS] above). If it does not match, do not install it yet, tell me what you found and let me decide. 2. Safety check: review the SKILL.md and scripts for anything suspicious (unexpected network calls, shell commands, credential harvesting). Beyond the fingerprint match in step 1, this repo should contain zero secrets in code or hardcoded anywhere, and no commands beyond the pinned extractor (npx [email protected] init --no-skills). Verify that holds here; STOP on any red flag and tell me. Confirm in particular the new layering rules from the file: a shipped design system is imported only when knowledge.sources names it, its entries carry their origin and the installed version stamp, repository entries win name collisions, UI guides are recorded as pointers and never copied, and guards.json is written by the team or imported from a layer — the extractor never touches it, and a team-authored guard rule is added only after the user confirms the exact pattern and its severity. 3. Install it as a skill: copy SKILL.md and its helper files into the agent's skills directory, in a folder named "om-ux-setup". 4. Verify with no network calls: frontmatter valid, files in place. 5. Report what was installed, where, and what I still need to do myself (e.g. point the agent at the repository to scan; decide whether to refresh an existing contract vs leave it; answer the 2-3 judgment-call questions; companion skills (om-ux-review-pr for PR review — including its --guard static check and --health repository modes, om-ux-shape for feature shaping, om-figma-design-with-ds for design-tool files, om-mockup-prototype for mockups) are separate collection members — install separately if needed). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.
Saved to your recent installs. Find it anytime on /connect.
🔒 Security, provenance and what this touches
- It can download and install software packages.
What the scan found — excerpts from the install prompt
- installs packages — Install "Extract a repo's design system into an executable UX contract (.uxproof/)" for me. It gives my agent · Installe-moi « Extraire le design system d'un repo en contrat UX exécutable (.uxproof/) ». Il donne à mon agen
The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Verification certificate
Dated facts about what was checked for this listing, and how. This is not a safety certification.
- Pinned version
- e320bd5 · 2026-10-21
- Sandbox install test
- Install ran in our test sandbox · Nov 11, 2026
- Creator identity
- Not verified
- Fingerprint watch
- Last checked · Nov 19, 2026 : no change detected in the files linked by this install prompt
- Install prompt analysis
- 2026-10-13 : installs packages
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Questions
How do I install a skill?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.