design-system-review
Skill0 starsChanged 3 months agoMITView on GitHub
Review a design system, UI library, token architecture, component inventory, Figma library, Storybook, or implemented product for consistency, coverage, accessibility, maintainability, adoption, and governance. Use when asked to audit components, tokens, spacing, typography, color, theming, variants, states, APIs, documentation, design-code parity, contribution, or migration. Produce evidence-backed system findings, a component and token remediation model, governance decisions, and a sequenced adoption plan. Do not use to police intentional product variation or to create a visual brand from scratch.
3 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- codeotter0201/design-system-skillstheir own version, first seen Nov 8, 2026
- Iceberg86-gb/labelhub-platformtheir own version, first seen Nov 8, 2026
- aditya-ariosity/ux-ui-skillsedited, first seen Sep 2, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- aditya-ariosity/ux-ui-skills★ 262% same
Wrote their own version · 2
- Iceberg86-gb/labelhub-platform★ 48same name
- codeotter0201/design-system-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- ai-product-designSkill
- case-study-writerSkill
- dashboard-redesignSkill
- handoff-to-devSkill
- ux-ui-auditSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 7 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 20
Do not equate visual inconsistency with system failure until intent and ownership are known.
- Distinctive: line 34
Read references/tokens-and-theming.md.
- Distinctive: line 38
Read references/component-contract.md.
- Distinctive: line 46
Read references/governance-and-adoption.md.
- Distinctive: line 61
Do not recommend one mega-component to eliminate every variation.
- Very specific: line 73
Use Critical, Major, Moderate, and Minor only when the organization already uses those terms.
- Distinctive: line 87
Avoid an unweighted score unless criteria and evidence support it.
- Distinctive: line 115
all consequential states are part of the component contract.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Design System Review
Review the system as a product and an operating model, not a screenshot library. A healthy system creates shared decisions, accessible defaults, predictable behavior, and a credible path from design to production.
1. Define the system boundary
Establish:
- products, platforms, brands, themes, locales, and teams served;
- available evidence: token files, component code, Storybook, Figma libraries, documentation, usage analytics, product screenshots, issue logs;
- source of truth for tokens, components, content, and release status;
- expected consumers and valid extension points;
- maturity stage: inventory, emerging library, adopted system, or multi-brand platform.
Do not equate visual inconsistency with system failure until intent and ownership are known.
2. Build an evidence inventory
Sample real product usage, not only pristine documentation. Create:
Pattern | Instances | Intended source | Observed variants | State gaps | Accessibility risk | Adoption signal | Owner
Trace representative components from semantic token to design asset to code to product. Record drift at the layer where it originates.
3. Review four layers
Foundations and tokens
Read references/tokens-and-theming.md. Inspect semantic naming, primitive separation, modes, themes, aliases, units, typography, motion, elevation, density, breakpoints, deprecation, and distribution.
Components and patterns
Read references/component-contract.md. Inspect anatomy, variants, states, content, behavior, semantics, keyboard support, responsiveness, composition, APIs, escape hatches, and test coverage.
Documentation and tooling
Inspect discoverability, examples with real content, do/don't rationale, design-code linkage, changelogs, migration guidance, release channels, and executable stories or tests.
Governance and adoption
Read references/governance-and-adoption.md. Inspect ownership, contribution, decision rights, review service levels, versioning, deprecation, exception handling, support, usage measurement, and funding.
4. Diagnose causes, not counts
Cluster symptoms into system causes such as:
- missing semantic layer;
- overloaded component API;
- absent state or content contract;
- inaccessible primitive;
- design-code source conflict;
- undocumented extension path;
- weak contribution and release process;
- migration cost with no product incentive.
Do not recommend one mega-component to eliminate every variation. Preserve legitimate domain differences and standardize the decisions that are truly shared.
5. Calibrate findings
Rate each finding by:
- product and accessibility impact;
- reach across instances and teams;
- recurrence and future drift risk;
- migration cost and dependency;
- confidence in evidence.
Use Critical, Major, Moderate, and Minor only when the organization already uses those terms; otherwise use Blocker, Systemic, Local, and Opportunity. Keep effort separate from impact.
6. Produce the review
System frame
State scope, maturity, evidence, exclusions, and source-of-truth assumptions.
Health summary
Use a table:
Layer | Health | Strong evidence | Main risk | Confidence
Avoid an unweighted score unless criteria and evidence support it.
Prioritized findings
For each:
Evidence | Product impact | System cause | Recommended decision | Migration path | Acceptance criteria
Token and component plan
Show proposed semantic layers, component contracts, additions, consolidations, deprecations, and intentional exceptions.
Governance plan
Define owners, contribution flow, release and deprecation rules, support, and adoption measures.
Sequence
- Stabilize: accessibility, broken contracts, conflicting sources, high-risk drift.
- Consolidate: semantic tokens, component variants, documentation, parity.
- Adopt: migrations, tooling, product integration, measurement.
- Evolve: new patterns, multi-brand needs, experiments.
Quality bar
- evidence includes real product instances;
- appearance, behavior, semantics, and API consistency are reviewed separately;
- tokens express meaning, not only raw values;
- all consequential states are part of the component contract;
- accessible defaults are built into primitives;
- exceptions have explicit owners and rationale;
- design and code have a declared authority and synchronization model;
- recommendations include migration and governance, not only cleanup.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

