🎉 Features
Design Token Usage Audit
New tool report-audit-token-usage — audits design-token usage across style files in two modes. validate: checks var() token references against the token dataset, flags unknown tokens, and suggests the closest match via edit-distance ("did you mean…?"), plus brand-specific availability warnings. overrides: detects token re-declarations in consumer files, broken down by mechanism and classification. Runs both by default (modes: ["all"]); supports excludePatterns, brandName, componentName, tokenPrefix, and saveAsFile.
Token Dataset Infrastructure
New configuration ds.generatedStylesRoot + ds.tokens.* loads a design-token dataset from your generated CSS (the engine behind the audit tool). Configurable discovery and modeling: filePattern (token CSS globs, default **/semantic.css), propertyPrefix (restrict extraction by prefix), scopeStrategy (flat or brand-theme — maps a {brand}/{theme}/ layout to scope metadata), and categoryInference (by-prefix, by-value, or none).
Storybook Story Parser
New tool get-ds-story-data — parses a DS component's Storybook .stories.ts into structured data: imports, argTypes, meta args, slots, selector style, form integration, and story definitions. Supports format (markdown/json), descriptionLength (short/full), and excludeTags (defaults to filtering docs-template showcase stories).
🐛 Bug Fixes
Multi-Prefix Validation — token validation now handles datasets using multiple prefixes (e.g. --semantic- and --ds-) instead of a single fixed prefix.
!important Detection — token overrides declared with !important are now correctly detected and reported.
Unknown Mechanism Fallback — overrides using an unrecognized mechanism no longer drop from the report; they fall back to a safe classification.