docs: note variables in ROADMAP automation row

ROADMAP.md

@@ -103,6 +103,7 @@ grade by behaviour against `plan/illustrator/`, not by whether a menu item exist
 - **Architecture:** 19+ crates with enforced layering (`cargo xtask layers`). Every action is a command (~400 engine + ~50 UI). Undo is unlimited via structural sharing. `command.batch` runs several commands as one transaction.
 - **Automation:**
   - Actions panel (record/playback, persisted), generic parameter dialogs for every "…" command.
+  - Variables (data merge): text and visibility variables bound to objects, datasets applied in one undo step (`variable.*`, `dataset.*`), with a Variables panel; every command, like the rest, is drivable over MCP and the CLI.
   - JSON-lines control channel with real egui pointer and keyboard injection.
   - MCP server with 25 tools (drawing, text, effects, Pathfinder, transforms, graphs, text wrap, export, screenshots, any command), attached to the running app or headless. Tool descriptions and replies agree with the engine: format lists come from its format table, `save_file` says it keeps the document's own format, `inspect_document` reports the paint type's characters show, and selection exports report the encoder's warnings. `--automation-read-root` / `--automation-write-root` (on `vectorcraft-cli mcp` and on `vectorcraft --control`) confine the files commands read and write to chosen folders, links followed, with the flags PhotoCraft takes (#832).
   - MCP protocol: five prompts (`prompts/list` / `prompts/get`: poster, icon set, recolour, trace-and-style, export set), `completion/complete` for prompt arguments and template variables read from the live catalogues (formats, effect ids, command ids, trace presets, swatch and object ids), four resource templates (`vectorcraft://object/{id}`, `//command/{id}`, `//effect/{id}`, `//swatch/{name}`) so an agent reads one thing instead of the whole document, and `logging/setLevel` with `notifications/message`. `window.panel` ids are case-insensitive (display labels work too), `vectorcraft://object/{id}` reads through `document.node {summary: true}` (the same reply bytes from one lookup instead of summarizing the whole document: a leaf read on a 2,000-object document drops from ~20 ms to ~0.06 ms), large containers slice with `document.node {depth?, childLimit?}` (truncated levels report `childCount`), `document.inspect` and the `inspect_document` tool slice the same way so a depth-0 read is the layer skeleton (~9.7 KB instead of ~2 MB on a 2,000-object document), `document.find` locates nodes by name/kind/text with ancestor paths, and the CLI sweeps the protocol surface over stdio against the compiled binary. `docs/mcp.md` has a Protocol section covering the capability table and what is deliberately absent (subscriptions, progress, cancellation, sampling, elicitation, Streamable HTTP — all of which need the transport to interleave reads).
@@ -241,7 +242,7 @@ about 1–1.5 agent-hours each, so the other rows (estimated on the older scale)
 | File formats | 6 | 85% | Affinity export, and the rest of Affinity import (effects, adjustments, brushes, master pages, live symbol instances), DWG (no open spec: DXF instead), PSD placement as layers (Photoshop Import Options; the merged image is done), EPS checked against more real files users send (Inkscape files open as vectors; other generators' constructs are tested with programs written from the PostScript Language Reference only; CID-keyed CJK type and `CCITTFaxDecode` images not read), legacy `.ai` files saved without the prolog that defines their operators (their own operator set isn't read yet), the rest of Illustrator's editing data (appearances and live effects, symbols, pattern fills, brushes, placed files: #637), the remaining fidelity polish; EPS, DXF, EMF/WMF, TIFF, BMP, Targa, PSD export, SVGZ, PDF security, PDF/X, PDF layers, presets and Affinity import are done | 6–10 (existing scope; Affinity export unestimated) |
 | Export for Screens, Asset Export, slices, Save for Web | 3 | 85% | polish only (all four are done) | 2–4 |
 | Print, colour management, separations, flattener | 3 | 80% | print preview fidelity (flattener presets in PDF and print, overprint in PDF and composite print, Print dialog, presets, PostScript, marks, separations, Print Tiling, print overprint/flattener/bitmap options and printer profile are done) | 1–2 |
-| Automation | 3 | 60% | Variables (data merge), scripting surface, batch | 8–12 |
+| Automation | 3 | 62% | scripting surface, batch (Variables: text/visibility variables bound to objects, datasets applied in one undo step, panel; image/graph kinds and dataset import open) | 6–10 |
 | UI chrome (panels, contextual Properties, workspaces, prefs) | 6 | 78% | macOS native menu dismissal fixed (M14.8); remaining: Variables, SVG Interactivity, Properties per context, Consolidate All Windows, floating panels outside the app window and docked into any group, context menus beyond the canvas (Links, Asset Export, CSS Properties, Attributes and the canvas context menu are done), about 17 Preferences options stored but not yet acted on (#394: Highlight Alternate Glyphs, Open Documents As Tabs, Real-time Drawing, hyphenation language, Appearance of Black, devices…) | 14–22 |
 | Libraries, Links, Package | 2 | 82% | the Libraries panel is local only (no cloud by design: libraries are files on this machine) and has no shared libraries, linked graphics or gradients and patterns as items yet; its graphics, colours and text styles (#745) and its search, groups and library export/import as `.vclibrary` files (#926) are done; Links and Package are done; placed VectorCraft documents are done (linked by default, vectors in every output, previews, Break Link, Expand, Edit Original, refresh on save) | 2–3 |
 | **Feature parity** | **103** | **~75%** | | **~210–330** |