Extract design tokens and publicly visible CSS information from any website
Strong parameter schemas and descriptions across all 9 tools. Tool names follow verb_noun convention (extract_*, get_*, analyze_*). Most parameters are well-documented with types, defaults, and constraints. However, output schemas are not explicitly documented in the source code, only inferred from tool descriptions. Error handling guidance is minimal. Tool descriptions are solid (100-200 chars) but could better explain WHEN to use each variant (extract vs extract_dtcg vs extract_tailwind). No tool annotations (readOnlyHint, destructiveHint) despite all being read-only operations.
Analyze WCAG contrast ratios between foreground and background colors.
Extract design tokens from a website URL. Returns comprehensive branding data including colors, typography, spacing, components, and more.
Extract design tokens and generate a brand guide PDF.
Extract design tokens and generate a DESIGN.md file documenting the design system.
Extract design tokens in W3C Design Tokens Community Group (DTCG) format from a website.
Extract design tokens and generate a self-contained HTML report from a website.
Extract design tokens and generate a shadcn/ui theme block with observed slots.
Extract design tokens and generate a Tailwind v4 @theme CSS file with observed values.
Output schemas not documented. Tool descriptions state what is returned (e.g., 'comprehensive branding data', 'DTCG format', 'HTML report') but do not specify the JSON structure, field names, or types. LLMs cannot plan downstream operations without knowing response shape.
No tool annotations despite all tools being read-only. The MCP spec supports readOnlyHint to signal safe tools. This helps agents reason about side effects and retry safety.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 67 | 2026-07-28+ | v2 |
Get the status of an asynchronous extraction job.
Seven extract_* tools are nearly identical in parameters and behavior, differing only in output format. Descriptions do not explain when to choose extract vs extract_dtcg vs extract_tailwind. LLMs may pick the wrong variant.
Error handling guidance absent. No description of what errors can occur, how to recover, or what the LLM should do if extraction fails (e.g., 'URL unreachable', 'timeout', 'no design tokens found'). Agents have no recovery path.
Async job pattern (sync=false returns job_id, get_job_status polls) is not documented in tool descriptions. LLMs may not understand they must call get_job_status to retrieve results, or may not know when a job is complete.