/design:new
Create a new multi-artboard canvas project via frontend-design — generic envelope adapted to .design/config.json. Default = --perfect (8 iter, full panel, target 4.5/5). Opt out via --quick or --no-critic. Opt out of the DS via --opt-out=palette|aesthetic|full.
| Property | Value |
|---|---|
| Command | /design:new |
| Category | daily (daily) |
| Argument hint | <Name> \"<brief>\" [--blank] [--from-annotations] [--fresh] [--component] [--mobile] [--quick | --no-critic] [--perfect-iter N] [--opt-out=palette|aesthetic|full] [--ds=<name>] |
| Source | plugins/design/commands/new.md |
Description
Create a new multi-artboard canvas project via frontend-design — generic envelope adapted to .design/config.json. Default = --perfect (8 iter, full panel, target 4.5/5). Opt out via --quick or --no-critic. Opt out of the DS via --opt-out=palette|aesthetic|full.
Invocation
/design:new <Name> \"<brief>\" [--blank] [--from-annotations] [--fresh] [--component] [--mobile] [--quick | --no-critic] [--perfect-iter N] [--opt-out=palette|aesthetic|full] [--ds=<name>]Summary
Creates a new multi-artboard canvas file at <designRoot>/<newCanvasDir>/<Name>.tsx via the frontend-design plugin. The generic envelope adapts to <repo>/.design/config.json (rootClass, themeDefault, tokensCssRel, …). The canvas envelope (DesignCanvas / DCSection / DCArtboard) is imported from the virtua
Source of truth
/design:import
Bring an existing design into Maude. `--reconstruct <image>` (T15, experimental) turns a Figma-frame PNG export into a real, token-styled canvas via a Bash-free vision-authoring + reality-check loop (DDR-174). Token files and brand material have their own dedicated entry points — see Notes below.
/design:photo
Edit a photo in the active canvas headlessly — parametric adjustments (brightness/contrast/saturation/exposure/hue/sepia/grayscale/invert), duotone, grain, pattern overlays, preset masks, and magic background removal — via the non-destructive PhotoEdit sidecar. Every edit is scriptable (no slider-clicking): parametric edits go straight to the /_api/photo-edit route; background removal runs client-side ML through a headless harness.