Loading ToolBench
MCP server exposing Quieto Tokens design token generation, contrast checking, and token system analysis.
Quieto MCP demonstrates solid tool design with clear naming, comprehensive descriptions, and well-structured schemas. All 7 tools follow verb_noun conventions (generate_*, check_*, map_*, bootstrap_*). Descriptions are detailed (100-300+ chars) and explain WHAT, WHEN, and WHAT-RETURNS. Input schemas use Zod with proper type constraints (enums for scale/base, string validation for colors). However, output schemas are not formally documented, responses are JSON.stringify'd without explicit schema declarations. Error handling is minimal; no recovery guidance or validation error messages visible. All tools are READ_ONLY (safe), but no tool annotations present. Parameter descriptions are thorough and include format hints (e.g., 'Hex color string (e.g. #4F46E5)').
Bootstrap a token system from an existing codebase's hardcoded styles instead of answering the quick-start prompts from scratch. Pass the contents of the project's stylesheets (.css/.scss/.sass/.less/.styl); Quieto tallies the colors, spacing, font families, and weights it finds and infers the seed inputs (brand color, additional hues by role, spacing base, type scale, fonts/weights, and light/dark themes). This is a seed-and-generate flow: by default it also runs the normal accessible-ramp pipeline to produce a full themed token system. Returns an inference summary (with warnings when it had to guess) so you can review and override before writing anything.
Check WCAG 2.1 contrast ratio between two hex colors. Returns ratio and pass/fail for AA and AAA at normal and large text sizes.
Generate a 10-step accessible color ramp (50-900) from a brand hex color. Returns primary and optional neutral ramps as DTCG-aligned tokens.
Generate a spacing primitive token scale from a base unit (4px or 8px). Returns DTCG-aligned spacing tokens.
Generate light and dark theme token collections. The dark theme inverts color steps to maintain contrast and vibrancy against dark surfaces.
Output schemas not formally documented. Tools return JSON.stringify'd objects without explicit schema declarations. LLMs cannot reliably parse or plan downstream calls without knowing field names, types, and structure.
No error handling or recovery guidance. Tools lack validation error messages, timeout handling, or actionable error responses. If a user passes an invalid hex color or malformed stylesheet, the LLM receives a raw exception with no guidance on how to fix it.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). While all tools are READ_ONLY, this is not declared in the schema. LLMs cannot infer safety properties and may avoid retrying on transient failures.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 63 | 2026-07-28+ | v2 |
Generate typography primitive tokens (font sizes, weights) from a scale preset. Returns DTCG-aligned tokens.
Generate semantic tokens (tier 2) from primitive inputs. Maps color, spacing, and typography to UI roles like 'background.primary', 'spacing.md', 'typography.body'.
bootstrap_from_codebase accepts raw stylesheet content as a parameter. No input sanitization visible against CSS injection or path traversal. If stylesheets are user-provided, this could be exploited.
No pagination or result limits documented. bootstrap_from_codebase could return very large token collections if processing a large codebase. No guidance on max results or chunking strategy.