plumb_fig_node

Read one node from a saved .fig file by its id. The headless counterpart to plumb_node — works with no Figma desktop, no plugin pairing, no FIGMA_TOKEN.

Inputs

Field	Type	Notes
figPath	string · required	Absolute path to a .fig file on disk.
id	string · required	Node id — sessionID:localID form. Get one from plumb_fig_outline.

Returns

{
  "source": "fig",
  "node": {
    "id":          "23:4567",
    "name":        "Home",
    "type":        "FRAME",
    "size":        { "w": 1440, "h": 1024 },
    "stackMode":   "VERTICAL",        // auto-layout direction, or undefined
    "cornerRadius": 12,
    "opacity":     1,
    "fillCount":   1,
    "strokeCount": 0,
    "text":        null               // populated for TEXT nodes
  },
  "next": "This MVP exposes basic metadata. …"
}

For TEXT nodes the text object carries characters, fontSize, fontFamily, and fontStyle.

When the agent should use it

• After plumb_fig_outline, to inspect a specific screen by id.
• Quick sanity-check that a .fig contains the node you expect.

MVP scope

This is the minimum useful surface for the .fig path. It returns metadata, not the full Plumb Design Spec yet — auto-layout → flexbox, deduped tokens, recursive children with depth-stable el handles are on the roadmap.

For the full PDS today, use plumb_node on the plugin or REST path. Use this tool when you need the metadata in headless / CI / offline contexts where the live paths aren't available.