figma-extraction-rules

name	Figma Extraction Rules
description	Rules for extracting and cleaning Figma design data
user-invocable	false

Figma Data Extraction Rules

Node Type Filtering

Keep (semantic nodes)

• FRAME - layout containers
• GROUP - logical groupings
• COMPONENT - reusable design elements
• INSTANCE - component instances
• COMPONENT_SET - variant groups
• TEXT - text content
• SECTION - top-level sections

Skip (visual-only nodes)

• VECTOR - paths/shapes
• RECTANGLE - decorative boxes
• ELLIPSE - circles/ovals
• LINE - line elements
• BOOLEAN_OPERATION - combined shapes
• STAR, POLYGON - decorative shapes

Depth Limiting

• Maximum depth: 8 levels
• Beyond 8 levels, collapse to summary

Repeated Element Collapsing

• If >3 children have identical names: collapse to "Name (×N)"
• Preserves one example of the pattern
• Reduces output size significantly

Data Cleaning

• Remove internal Figma IDs where not needed for reference
• Strip style properties (fills, strokes, effects)
• Keep: id, name, type, children, characters (for TEXT)
• Preserve frame dimensions for layout analysis

URL Parsing

• Format: figma.com/design/:fileKey/:fileName?node-id=:nodeId
• Convert nodeId dashes to colons: 123-456 → 123:456
• Branch URLs: use branchKey as fileKey