Skip to main content
Glama

Designesy

by io.github.LE-VAI
OverviewDiscussionsAdmin

Generate build-contract bundle

designesy_guardrails
Read-onlyIdempotent

Generate a frozen build-contract bundle for AI coding agents from any design system URL (the product layer). Ingests a site, extracts its :root tokens, and emits 6 outputs: (1) DTCG-format token file, (2) Stylelint config generated from token values, (3) AGENTS.md-format rules with token allowlist, (4) component contract with allowed prop patterns, (5) anti-pattern documentation, (6) DESIGN.md file (Google open spec, google-labs-code/design.md), the de-facto AI-readable design-context standard: YAML front matter plus a markdown body. Use this when you need to turn a design system into the file AI agents read and the lint that enforces it. When NOT to use: for design-contract scoring, use designesy_score; for token-file validation, use designesy_tokens_score; for drift detection, use designesy_drift_score. Executable: fetches the URL, extracts CSS + :root custom properties, generates the bundle. No browser needed. The full bundle grows with the number of tokens a site declares: about 7 KB to about 530 KB of JSON on the 26 sites measured (about 110 KB for designesy.org). To trim it, pass parts, a list of bundle file names such as ["designMd"] or ["tokens", "lintConfig"]: the result keeps score, grade, counts and checks, and its bundle holds only those files, in the order asked. Valid parts: tokens, lintConfig, agentRules, componentContract, antiPatterns, designMd. A single part can still be large on a site with many tokens. An unknown name returns an error that lists the valid parts. Returns JSON: { ok, url, score (0-100, emission completeness), grade, pass, warn, fail, total, tokensExtracted, bundle: { tokens, lintConfig, agentRules, componentContract, antiPatterns, designMd }, checks[{id, item, category, status, detail}] }; with parts it adds parts, and bundle holds only the named files. Results cached ~24h per URL.

Input Schema

NameRequiredDescriptionDefault
urlNoURL to generate guardrails for. Defaults to https://www.designesy.org/ if not provided.
partsNoOptional: the bundle files to return, by name. Omit for the whole bundle. Valid parts: tokens, lintConfig, agentRules, componentContract, antiPatterns, designMd.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changed
    • addedInput schema / properties / parts
      Added value: +{
+  "description": "Optional: the bundle files to return, by name. Omit for the whole bundle. Valid parts: tokens, lintConfig, agentRules, componentContract, antiPatterns, designMd.",
+  "items": {
+    "type": "string"
+  },
+  "type": "array"
+}
  2. Changed2 schema fields changed
    • changedInput schema / $schema
      Previous value: -"http://json-schema.org/draft-07/schema#"New value: +"https://json-schema.org/draft/2020-12/schema"
    • removedInput schema / additionalProperties
      Removed value: -false
  3. Added

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations cover the safety profile (readOnly, idempotent, openWorld, non-destructive), and the description adds genuinely useful context beyond them: it fetches the URL and extracts CSS/:root properties with no browser, bundle size range (7KB–530KB), ~24h per-URL caching, deterministic ordering of returned parts, and that an unknown part name returns an error listing valid names.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with purpose and the six outputs before the routing guidance, which is the right ordering. However the same six part names are enumerated three times (outputs list, 'Valid parts', and the return JSON), which is noticeable redundancy in an otherwise dense block.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema, the description fully carries the return-value burden by spelling out the JSON shape (ok, url, score, grade, counts, bundle, checks) and how parts changes it. For a complex multi-file generator, an agent has everything needed to call it correctly and interpret the result.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, but the description adds real meaning: the url defaults to designesy.org, parts trims the bundle to named files (preserving score/grade/counts/checks), part order is honored, and bad names produce an informative error. It stops short of restating per-parameter types, hence 4 rather than 5.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource ('Generate a frozen build-contract bundle for AI coding agents from any design system URL') and enumerates the six concrete outputs. An agent can distinguish it from every sibling without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Provides explicit when-to-use ('turn a design system into the file AI agents read and the lint that enforces it') and an explicit when-NOT-to-use list routing to designesy_score, designesy_tokens_score, and designesy_drift_score by condition. Nothing is left to inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Designesy toolsInspect
Try in Browser
Shareable, User-Owned Context BackendPeople, applications, and agents read, write, and coordinate together in one context lake. Reach it over MCP, a CLI, a Python SDK, or REST.fulcradynamics.comAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources