I’ve been trying a new workflow for auditing correct token usage of components (through skills in programs like Cursor), however I encounter an issue whenever I’m trying to audit actual Figma components instead of un-componentized frames. When I audit a variant of a component (other than the base component):get_design_context returns code with the base component's colors (e.g., surface/neutral/primary, border/neutral/secondary)
get_variable_defs returns the variant's actual variables (e.g., surface/neutral/primary-hover, border/neutral/secondary-hover, icon/neutral/tertiary-hover)
This creates a mismatch between the generated code and the actual design, making it impossible to programmatically audit or generate accurate code for component variants. The only way to verify correct token usage is manual inspection in Figma.
Is this a known limitation? Is there a way for the generated code of get_design_context to return the variant’s variables instead of the base component? This would solve it.