Skip to main content
Yanked

This release has been yanked by its maintainers, and will be ignored by installers, except when explicitly specified.
Consider using release 1.13.10 instead.
Reason given by maintainers: Erroneous early upload — malformed version. Use 1.11.x instead.

designesy-mcp

PyPI version Python 3.10+ License: MIT MCP

One-click install:

Install in VS Code Install in VS Code Insiders Install in Cursor Install Remote (HTTP)

A read-only stdio MCP server exposing designesy.org's design-intelligence infrastructure as native agent tools.

Zero external dependencies. Pure Python stdlib. Implements the Model Context Protocol JSON-RPC 2.0 over stdio.


Quick start (one command)

uvx designesy-mcp

That's it. uvx fetches the package from PyPI, creates an ephemeral environment, and launches the stdio MCP server. No virtualenv, no pip install, no git clone. The server is ready to speak JSON-RPC 2.0 on stdin/stdout immediately.

Don't have uv? Install it once: curl -LsSf https://astral.sh/uv/install.sh | sh (macOS/Linux) or powershell -c "irm https://astral.sh/uv/install.ps1 | iex" (Windows). Or use pipx run designesy-mcp as an equivalent one-liner.

MCP client config

{
  "mcpServers": {
    "designesy": {
      "command": "uvx",
      "args": ["designesy-mcp"]
    }
  }
}

pip install (traditional)

pip install designesy-mcp
{
  "mcpServers": {
    "designesy": {
      "command": "designesy-mcp"
    }
  }
}

python -m (module form)

{
  "mcpServers": {
    "designesy": {
      "command": "python",
      "args": ["-m", "designesy_mcp_server"]
    }
  }
}

No arguments needed. The server speaks JSON-RPC 2.0 over stdin/stdout.

Tools (17)

The server exposes 17 tools, all fetched live from https://www.designesy.org/:

Read-only discovery

Tool What it does
designesy_catalog Get the 12-package catalog (versions, URLs, statuses) from /open.json
designesy_contract Get the full design-system contract v0.4.0 (tokens, motion, acoustic, takt, cadence, typography, components, verification, open tensions) — or a filtered section
designesy_design_review Get the Design Review kit (8 dimensions, agent prompt, output format, verification checklist)
designesy_skill_md Get the agent-skill-format export (SKILL.md) with behavioral rules, tokens, anti-patterns
designesy_agent_json Get the agent discovery document (.well-known/agent.json) — identity, authority, ingest protocol
designesy_llms_txt Get the short agent brief (/llms.txt)
designesy_llms_full_txt Get the full agent brief (/llms-full.txt) with ingest protocol, all packages, paste-ready prompt

Executable verification

Tool What it does
designesy_score Run the 40-check contract verification against a live URL. Fetches HTML + CSS, parses :root custom properties, returns PASS/FAIL/WARN/SKIP per check with an overall score, letter grade, and per-category breakdown. Supports 4 emission formats: designesy (default), canonical (review-findings.json schema), review (jakubkrehel markdown), google (@google/design.md JSON).
designesy_tokens_score Validate a design token file against the W3C Design Tokens Community Group (DTCG) 2025.10 format. 10 checks (t01–t10).
designesy_a11y_score Get the WCAG 2.2 AA accessibility verification framework (11 checks, a01–a11) + a Playwright/axe-core script template for local execution.
designesy_motion_score Validate a Lottie animation file against Lottie spec v1.0.1 + the Designesy 10 Non-Negotiable Motion Standards. 10 checks (m01–m10).

Executable engines (new in v1.10.0)

Tool What it does
designesy_drift_score 12-check AI-drift radar — detects token fabrication, within-session drift, between-session amnesia, and silent breaking changes.
designesy_readiness_score 10-check AI readiness probe — tests for DTCG tokens, llms.txt, agent.json, MCP endpoint, DESIGN.md, sitemap, robots, OG meta.
designesy_guardrails Generate a frozen build-contract bundle: DTCG tokens, Stylelint config, AGENTS.md rules, component contract, anti-patterns, DESIGN.md.
designesy_monitor_score Continuous drift governance — 10 monitor checks with history deltas, trend slope, and email alerts via Resend.
designesy_compare Diff two design systems from live URLs — 8-dimension structured diff (added, removed, renamed, value-changed, scale, contrast, structure, score).
designesy_report Composite synthesis — fires score + drift + readiness in parallel, computes weighted composite grade. The most shareable surface.

Resources (7)

The server also exposes 7 MCP resources (read-only URIs):

URI Content
designesy://open Package catalog (JSON)
designesy://contract Full design-system contract (JSON)
designesy://kit/design-review Design Review kit (JSON)
designesy://skill SKILL.md agent-skill export (Markdown)
designesy://agent Agent discovery document (JSON)
designesy://llms Short agent brief (text)
designesy://llms-full Full agent brief (text)

The 40-check verification engine

designesy_score runs 40 deterministic checks across 13 weighted categories:

Category Weight What it measures
cadence 18 Typography rhythm — line-height, font-synthesis, text-underline-position, skip-ink
accessibility 15 WCAG 2.2 primitives — reduced-motion, forced-colors, AI disclosure, focus-visible
semantic 12 Token architecture — :root custom properties, no raw hex, semantic naming
motion 10 Motion hygiene — duration tokens, easing tokens, reduced-motion blocks
tokens 9 DTCG 2025.10 conformance — $type, $value, $description, colorSpace
takt 8 Timing discipline — transition bands, animation hierarchy
poise 7 Composure — viewport overflow, scroll behavior, print styles
identity 6 Brand coherence — title, meta description, favicon, og tags
interaction 6 Interaction primitives — hover states, press feedback, disabled states
performance 6 Core Web Vitals readiness — preload, font-display, render-blocking
responsive 3 Responsive primitives — viewport meta, container queries
security 5 Security headers — CSP, X-Content-Type-Options, referrer policy
spec 4 Spec conformance — lang attr, charset, doctype

No LLM. No roast. The same engine scores designesy.org itself — in public, at 95.3% A.

Standards positioning

DTCG 2025.10 — the spec went stable

The W3C Design Tokens Community Group published the spec's first stable version on Oct 28, 2025 — the Final Community Group Report, classified as a Candidate Recommendation and considered stable. 24+ organizations back it (Adobe, Google, Meta, Figma, Amazon, Microsoft, Shopify, Sketch, Framer). 84% of teams now use design tokens (2026, up from 56% YoY).

designesy_tokens_score validates against this stable spec. Every team adopting DTCG 2025.10 needs a validator — designesy is it.

Motion tokens — the spec's blind spot

The DTCG 2025.10 spec leaves motion tokens as a second-class citizen — there is no standard for motion token structure, reduced-motion markers, or animation accessibility. The 2026 State of Design Systems field report confirms this is the remaining friction point.

designesy_motion_score fills this gap. It validates Lottie files against the Lottie spec v1.0.1 AND the Designesy §16 Ten Non-Negotiable Motion Standards — the only validator that checks both structural well-formedness and accessibility (reduced-motion markers, no deprecated versions). It is the verification layer for the spec's known blind spot.

Contract vs. opinion

No competitor does contract-based deterministic scoring. Lighthouse is weighted heuristics. axe-core is rule violations. securityheaders.com is a single dimension. Designesy's 40-check contract-bound 0-100 score across 7 dimensions (tokens, motion, accessibility, cadence, takt, typography, copywriting) has no direct analog.

Caching

All responses are cached with a 5-minute TTL. The server only fetches public, machine-readable exports from designesy.org via HTTPS. It does not read local files, credentials, or source roots.

Safety

Read-only. This server never writes anywhere. It does not execute code, mutate files, or access credentials.

Provenance

All data is fetched live from:

  • https://www.designesy.org/open.json
  • https://www.designesy.org/contracts/design-system.json
  • https://www.designesy.org/kits/design-review.json
  • https://www.designesy.org/contracts/skill
  • https://www.designesy.org/.well-known/agent.json
  • https://www.designesy.org/llms.txt
  • https://www.designesy.org/llms-full.txt

License

MIT

Metadata

Release files for designesy-mcp 1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for designesy-mcp 1
File Size Uploaded Provenance Actions
designesy_mcp-1.tar.gz 45.6 kB Without provenance Details for designesy_mcp-1.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for designesy-mcp 1
File Interpreter ABI Platform Actions
designesy_mcp-1-py3-none-any.whl Python 3 none any Details for designesy_mcp-1-py3-none-any.whl

Total release size: 93.5 kB

Release files / designesy_mcp-1.tar.gz

Download URL designesy_mcp-1.tar.gz
Size 45.6 kB
Tags Source
SHA-256 checksum
How to use checksums
6036bcfc7c1e815254a13eace8fbf4bc76255e2cb9010617d69a1b4b19a7eb49
BLAKE2b-256 checksum
How to use checksums
0e5329f6f6f2760cecd960b250d97fe6161da203538c18643c8da1ed864d9ca4
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.12.13

Release files / designesy_mcp-1-py3-none-any.whl

Download URL designesy_mcp-1-py3-none-any.whl
Size 47.9 kB
Tags Python 3
SHA-256 checksum
How to use checksums
bc12fbdcf63702d7dd0186de3406d8eec1b3e1ea3986b4f6652b87d921fb032a
BLAKE2b-256 checksum
How to use checksums
e0aa020cd27d980ca228e5399ad35db4a00ccbbf245d9e55bb6ecfdaaa108304
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.12.13

Security for designesy-mcp 1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

1 This release

2 release files No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page