| name | create-property |
| description | Generate a visual property annotation in Figma showing each configurable property axis with component instance previews. Use when the user mentions "property", "properties", "property annotation", "create property", or wants to document a component's configurable properties visually. |
Create Property Annotation
Generate a visual property annotation directly in Figma — one exhibit per variant axis and boolean toggle, each showing the available options as component instances with a summary table.
Execution contract (read first).
- This file is instructions to RUN, not a document to edit. Invoking the skill = render the property annotation into Figma from the input
.md.
- Never edit this
SKILL.md or any other skill file in response, even if one is open or focused in the editor. Modify a skill only when the user explicitly asks to change the skill itself.
- The input component
.md is a READ-ONLY source of truth. Never edit, append to, or "add a section" to it. The only artifact this skill produces is the Figma annotation. When the user asks to "create/add a section," "show," or "include" something, render it in the Figma annotation, never as an edit to the .md.
- Never call
AskQuestion, request confirmation, or pause for input (including before Figma writes, the expected output). On ambiguity, pick the most defensible option and continue.
- Only two legal stops: (a) Step 0 fail-fast when no
.md resolves; (b) one-line abort if the Figma MCP connection is dead.
MCP Adapter
Read uspecs.config.json → mcpProvider. Follow the matching column for every MCP call in this skill.
| Operation | figma-console | figma-mcp |
|---|
| Verify connection | figma_get_status | Skip — implicit. If first use_figma call fails, guide user to check MCP setup. |
| Navigate to file | figma_navigate with URL | Extract fileKey from URL (figma.com/design/:fileKey/...). No navigate needed. |
| Take screenshot | figma_take_screenshot | get_screenshot with fileKey + nodeId |
| Execute Plugin JS | figma_execute with code | use_figma with fileKey, code, description. Core logic is identical — see font loading note below for the one API difference (getRangeAllFontNames → tn.fontName). |
| Search components | figma_search_components | search_design_system with query + fileKey + includeComponents: true |
| Get file/component data | figma_get_file_data / figma_get_component | get_metadata or get_design_context with fileKey + nodeId |
| Get variables (file-wide) | figma_get_variables | use_figma script: return await figma.variables.getLocalVariableCollectionsAsync(); |
| Get token values | figma_get_token_values | use_figma script reading variable values per mode/collection |
| Get styles | figma_get_styles | search_design_system with includeStyles: true, or use_figma: return figma.getLocalPaintStyles(); |
| Get selection | figma_get_selection | use_figma script: return figma.currentPage.selection.map(n => ({id: n.id, name: n.name, type: n.type})); |
figma-mcp requires fileKey on every call. Extract it once from the user's Figma URL at the start of the workflow. For branch URLs (figma.com/design/:fileKey/branch/:branchKey/:fileName), use :branchKey as the fileKey.
figma-mcp page context: use_figma resets figma.currentPage to the first page on every call. When a script accesses a node from a previous step via getNodeByIdAsync(ID), the page content may not be loaded — findAll, findOne, and characters will fail with TypeError until the page is activated. Insert this page-loading block at the start of every script that references a previously-created node:
const pages = figma.root.children;
const targetPage = pages.find(p => p.name === '__PAGE_NAME__');
if (targetPage) await figma.setCurrentPageAsync(targetPage);
Replace __PAGE_NAME__ with the actual page name (determined during Step 7 when the template is placed). This loads the page content so child nodes are accessible.
figma-mcp font loading: getRangeAllFontNames is not available in the use_figma sandbox and will throw TypeError. Replace it with tn.fontName (returns { family, style } for single-font text, or figma.mixed for mixed-font text). findAll and findOne work normally after setCurrentPageAsync — they do not need replacement.
Replace the font-collection loop in every script from:
const fonts = tn.getRangeAllFontNames(0, tn.characters.length);
for (const f of fonts) {
const key = f.family + '|' + f.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(f); }
}
to:
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
And add .catch(() => {}) to the batch load: await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
Inputs Expected
- Component
.md spec (required, user-provided path) — the source-of-truth component spec produced by {{skill:create-component-md}}. The user tells you where this .md lives — use the exact path they provide; the .md may live anywhere. This skill rebuilds its property model from the .md's render-meta block; it does NOT re-extract the property surface from Figma. fileKey, nodeId, and compSetNodeId come from render-meta, never from a Figma link.
- Figma link (optional) — placement hint only (which page/frame to drop the rendered annotation on, including a cross-file destination) and the target for the two whitelisted minimal live reads in Steps 4a/4b. Never the source of the property contract.
Asymmetry from the other consumers. Unlike the Structure / Color / Voice consumers, the .md has no dedicated "Property" body section to copy. So create-property takes identity only from render-meta (which lets it skip extraction) and still authors its own property model, normalization, and exhibit plans with light in-memory reasoning. It does not parse a Property section body — there isn't one.
There is no screenshot-only path and no component-link extraction path. Without the component .md there is nothing to build from — see Step 0's fail-fast contract.
Workflow
Copy this checklist and update as you progress:
Task Progress:
- [ ] Step 0: Require the component `.md`; parse render-meta. FAIL FAST if missing.
- [ ] Step 1: Read instruction file (only as needed for normalization / exhibit-planning reasoning — NOT for re-extraction)
- [ ] Step 2: Verify MCP connection
- [ ] Step 3: Read template key from uspecs.config.json
- [ ] Step 4: Rebuild the property model from render-meta (variant axes, booleans, slots, instance-swaps, constitutive sub-components) — NO extraction
- [ ] Step 4a: WHITELISTED LIVE READ #2 — bounded variant-gated-boolean scan for requiredVariantOverrides
- [ ] Step 4b: WHITELISTED LIVE READ #1 — variable-collection lookup for the 6c variable-mode exhibit
- [ ] Step 4d: Normalize child properties IN MEMORY (coupled / unified slot / sibling boolean grouping) — no Figma reads
- [ ] Step 4e: Exhibit planning + context axis + briefDescription IN MEMORY — no Figma reads
- [ ] Step 5: Re-read instruction file (Pre-Render Validation Checklist, Common Mistakes, Do NOT) and audit
- [ ] Step 6: Navigate to destination (if different file)
- [ ] Step 7: Import and detach the Property template
- [ ] Step 8: Fill header fields
- [ ] Step 9: Build property exhibits with component instances
- [ ] Step 10: Visual validation
- [ ] Step 11: Completion link (reports sourceHash from render-meta)
Step 0: Require the component .md and parse render-meta (fail fast)
This skill is a consumer of the .md source of truth. It does not re-extract the property surface from Figma — extract-api/create-component-md already captured the component's variant axes, booleans, slots, instance-swaps, and constitutive sub-components into the .md's render-meta block. Your job is to rebuild the property model from render-meta and author the exhibits.
Asymmetry from the other consumers. The .md has NO dedicated "Property" body section to parse. So this skill gets identity only from render-meta (to skip extraction) and still authors its own property model + exhibit plans with light in-memory reasoning. There is no Property section body — render-meta is the only thing you read from the .md.
-
Resolve the .md path. Use the exact path the user gave, else an attached or open .md in context. The .md may live anywhere; do NOT invent or guess a path. If neither resolves to an existing file, abort per item 2. Never pause to ask the user which file to use.
-
Require the file. If no file exists at the resolved .md path, abort immediately with this exact single-line diagnostic and stop — do NOT fall back to extraction:
This skill requires the component's Markdown .md spec (produced by create-component-md). Provide the path to it. (create-component-md needs a _base.json from the uSpec Extract plugin.)
-
Parse the render-meta block (the fenced JSON between <!-- render-meta:start v=1 --> and <!-- render-meta:end -->; schema in references/component-md/agent-component-md-instruction.md § RENDER_META_JSON). Capture:
component = { componentName, compSetNodeId, isComponentSet }.variantAxes — { <axisName>: [<option>, …] }.variantAxesDefaults — { <axisName>: <default> }.booleanDefs[] — { key, default, associatedLayerName, associatedLayerId }. Each key is the raw component-property key setProperties expects.propertyDefs — keyed by raw key; each { type, default?, values?, preferredComponentKey?, associatedLayerName?, associatedLayerId? }. type ∈ VARIANT | BOOLEAN | INSTANCE_SWAP | SLOT | TEXT | NUMBER.slotContents[] — { slotName, slotNodeType, preferredComponents: [{ componentKey, componentName, componentId, componentSetId, isComponentSet, variantAxes, booleanDefs }] }.subComponents[] — { name, mainComponentName, subCompSetId, subCompVariantAxes, subCompVariantAxesDefaults, booleanOverrides }.fileKey, nodeId — for template placement and the Step 11 completion link.sourceHash — record it; the Step 11 completion footer reports it so downstream readers can detect drift between this render and the _base.json that produced the .md.
-
Rebuild the property model in memory — see Step 4. No Figma reads.
FORBIDDEN — do NOT re-extract. When the component .md is present (it always is past Step 0), you MUST NOT run the legacy extraction/tree-walk. Specifically:
- The old Step 4 extraction script — the
node.componentPropertyDefinitions walk that rebuilt variantAxes, booleanProps, instanceSwapProps, and slotProps from scratch — is deleted. It does not exist in this skill anymore. The property surface comes from render-meta (Step 4).
- The old Step 4a full variant-gating script (the one that re-enumerated every
boolDef and every variantAxis from propDefs and walked every variant child for every boolean) is deleted and replaced by the bounded Step 4a scan below — which is whitelisted live read #2.
- The old Step 4c child-discovery walk (
walkForInstances recursing the default variant for nested instances + the boolean-linkage resolver) is deleted. Constitutive child chapters now come from render-meta.subComponents (Step 4). Do NOT run any figma_execute / use_figma script that walks the component tree to rediscover child components or rebuild their property surface.
- The ONLY Figma reads this skill makes beyond the render scripts are the two whitelisted minimal live reads, both bounded:
- Variable-collection lookup (Step 4b) — resolve
COLLECTION_ID / MODE_ID / MODES_JSON for the 6c variable-mode exhibit by listing variable collections only. No component tree walk; no property re-enumeration.
- Variant-gated-boolean scan (Step 4a) — a small, bounded scan of only the booleans already in the model (from
render-meta.booleanDefs) to learn which ones are only meaningful under specific variant values (requiredVariantOverrides). It does NOT rebuild propertyDefs / variantAxes / slotProps and does NOT discover new properties.
Everything else (variant axes, boolean defs, slots, instance-swaps, constitutive sub-components) is read from render-meta and consumed verbatim. Normalization (Step 4d) and exhibit planning (Step 4e) run entirely in memory.
Step 1: Read Instructions (only as needed)
Read agent-property-instruction.md when you need it for the Data Validation, Exhibit Planning, Pre-Render Validation Checklist, Common Mistakes, and Do NOT reasoning that Steps 4d/4e/5 perform over the render-meta-derived model. Do NOT use it as a prompt to re-extract — the property surface is supplied by render-meta (Step 4).
Step 2: Verify MCP Connection
Read mcpProvider from uspecs.config.json to determine which Figma MCP to use.
If figma-console:
figma_get_status — Confirm Desktop Bridge plugin is active- If connection fails: "Please open Figma Desktop and run the Desktop Bridge plugin. Then try again."
If figma-mcp:
- Connection is verified implicitly on the first
use_figma call. No explicit check needed.
- If the first call fails: "Please verify your FIGMA_API_KEY is set correctly in your MCP configuration."
Step 3: Read Template Key
Read the file uspecs.config.json and extract:
- The
propertyOverview value from the templateKeys object → save as PROPERTY_TEMPLATE_KEY
- The
fontFamily value → save as FONT_FAMILY (default to Inter if not set)
If the template key is empty, tell the user:
The property template key is not configured. Run {{skill:firstrun}} with your Figma template library link first.
Step 4: Rebuild the property model from render-meta (no extraction)
Everything the normalization, exhibit planner, and render scripts need about the property surface comes from the render-meta block parsed in Step 0 — there is no extraction call here (see the Step 0 FORBIDDEN directive). Rebuild this in-memory model from render-meta (the field names below are exactly what the render scripts in Step 9 expect):
componentName / compSetNodeId / isComponentSet — from render-meta.component (componentName, compSetNodeId, isComponentSet). compSetNodeId is the COMP_SET_NODE_ID every render script consumes.variantAxes — reshape render-meta.variantAxes ({ axisName: [options] }) joined with render-meta.variantAxesDefaults ({ axisName: default }) into [{ name, options, defaultValue }], one entry per axis.defaultProps — render-meta.variantAxesDefaults verbatim ({ <axisName>: <default> }). This is the DEFAULT_PROPS baseline for 6a/6a-ctx/6a-matrix.booleanProps — one entry per render-meta.booleanDefs[]: { name: key.split('#')[0], defaultValue: <default>, associatedLayer: associatedLayerName, rawKey: key, controlsSlot, slotPreferredNames }. Carry the raw key as rawKey — setProperties() needs the raw-key form (cross-check propertyDefs[key].type === "BOOLEAN" when in doubt). Derive the slot fields in memory from render-meta (no Figma read): controlsSlot = true when associatedLayerName matches a render-meta.slotContents[].slotName; slotPreferredNames = <that slot's preferredComponents[].componentName> (else false / []). These feed the 6b "Controls slot" description.instanceSwapProps — one entry per render-meta.propertyDefs entry whose type === "INSTANCE_SWAP": { name: rawKey.split('#')[0], defaultValue: def.default, rawKey }.slotProps — one entry per render-meta.slotContents[]: { name: slotName, preferredInstances: preferredComponents.map(p => ({ componentKey: p.componentKey, componentName: p.componentName, componentId: p.componentId })), rawKey: <matching propertyDefs SLOT key, when present> }.childComponents (constitutive sub-component chapters) — from render-meta.subComponents[], constitutive sets only (see the SCOPE LIMIT below).
SCOPE LIMIT — constitutive sub-component-SETS only. Only entries in render-meta.subComponents with a non-null subCompSetId get full child chapters (rendered in 6e/6f/6g). For each kept entry, map to the childComponents shape the render scripts expect:
name = subComponents[].name; mainComponentName = subComponents[].mainComponentName.mainComponentSetId = subCompSetId; mainComponentId = null; isComponentSet = true.variantAxes = reshape subCompVariantAxes (joined with subCompVariantAxesDefaults) into [{ name, options, defaultValue }].booleanProps = the child booleans render-meta carried for this sub-component (derive from subComponents[].booleanOverrides keys when present; otherwise []).visible = true unless a controlling boolean is matched (below).controllingBooleanName / controllingBooleanRawKey = the parent boolean whose associatedLayerName (from booleanProps/render-meta.booleanDefs) matches this sub-component's name (light, in-memory name-match; fall back to normalized-name containment). null when no parent boolean controls it.
Collect the matched controlling-boolean names into a controllingBooleanNames array (Step 4d / 6b consume it).
DEGRADE GRACEFULLY — do NOT re-extract. Non-constitutive nested instances and child boolean/instance-swap surfaces that are NOT present in render-meta.subComponents (because the producer did not capture them as constitutive sets, or carried only booleanOverrides) may not get full chapters. Render what render-meta provides; when a child surface is incomplete, add a short note in the affected chapter (or the spec header) that the deeper surface lives in the child's own ./{slug}.md — do not run a live tree walk to rediscover it.
Save the rebuilt model — you will use it in subsequent steps.
Step 4a: Variant-gated-boolean scan (WHITELISTED LIVE READ #2 — bounded)
This is one of the only two Figma reads this skill makes beyond rendering. render-meta tells you which booleans exist and the layer each controls (booleanDefs[].associatedLayerName / associatedLayerId), but it does NOT record which variant a boolean's layer only appears under. A small, bounded scan resolves that so 6b can pick the right base variant. Example: a "Dismiss button" boolean may only control a layer that exists in the Behavior=Interactive variant, not in Behavior=Static; when the default variant lacks the target layer, toggling the boolean produces identical-looking previews.
Bound. Scan ONLY the booleans already in the model (from render-meta.booleanDefs), checking each one's layer against the default variant and — only when absent — across the variant children to find the minimal override. This is NOT a property re-extraction: it does not rebuild propertyDefs, variantAxes, slotProps, or booleanProps, and it does not discover new properties.
Run this bounded scan via figma_execute. Replace TARGET_NODE_ID with render-meta.component.compSetNodeId and __BOOLEAN_DEFS_JSON__ with the render-meta.booleanDefs[] array ({ key, default, associatedLayerName, associatedLayerId }):
const TARGET_NODE_ID = '__NODE_ID__';
const BOOLEAN_DEFS = __BOOLEAN_DEFS_JSON__;
const node = await figma.getNodeByIdAsync(TARGET_NODE_ID);
if (!node || node.type !== 'COMPONENT_SET') {
return { skip: true, reason: 'Not a component set — no variant gating possible', interpretedBooleans: [] };
}
const defaultVariant = node.defaultVariant || node.children[0];
const defaultVProps = defaultVariant.variantProperties || {};
const interpretedBooleans = [];
for (const bd of BOOLEAN_DEFS) {
const rawKey = bd.key || '';
const nodeIdSuffix = rawKey.split('#')[1] || null;
const result = { name: rawKey.split('#')[0], requiredVariantOverrides: null, layerName: bd.associatedLayerName || null };
if (!nodeIdSuffix) { interpretedBooleans.push(result); continue; }
let layerInDefault = null;
try {
const lid = defaultVariant.id.split(';')[0] + ';' + nodeIdSuffix;
const ln = await figma.getNodeByIdAsync(lid);
layerInDefault = ln ? ln.name : null;
} catch {}
if (layerInDefault) { result.layerName = layerInDefault; interpretedBooleans.push(result); continue; }
for (const child of node.children) {
const vp = child.variantProperties || {};
try {
const lid = child.id.split(';')[0] + ';' + nodeIdSuffix;
const ln = await figma.getNodeByIdAsync(lid);
if (ln) {
const diffAxis = {};
for (const [k, v] of Object.entries(vp)) {
if (defaultVProps[k] !== v) diffAxis[k] = v;
}
result.requiredVariantOverrides = diffAxis;
result.layerName = ln.name;
break;
}
} catch {}
}
interpretedBooleans.push(result);
}
return { interpretedBooleans };
How the agent should use this data:
The scan returns an interpretedBooleans array — one entry per boolean from render-meta.booleanDefs. Each entry contains:
name: the boolean's clean namerequiredVariantOverrides: an object like { "Behavior": "Interactive" } if the boolean is variant-gated, or null if it works on the default variantlayerName: the resolved layer name
For each boolean in interpretedBooleans:
requiredVariantOverrides === null — No action needed. The boolean works on the default variant. Render normally in 6b.requiredVariantOverrides is an object — The boolean is variant-gated. Store the requiredVariantOverrides on the boolean entry from Step 4's booleanProps. In 6b, use these overrides when looking up the base variant for instance creation. The description should note the dependency (e.g., "Requires Behavior = Interactive").
No AI reasoning is needed — the bounded scan has already resolved which booleans are variant-gated and what overrides they require. This scan is whitelisted live read #2; do not extend it into a full property re-extraction.
Step 4b: Variable-collection lookup (WHITELISTED LIVE READ #1 — bounded)
Some component properties (e.g., shape, density) are controlled via Figma variable modes at the container level, not per-instance. These never appear in componentPropertyDefinitions (and so are not part of the render-meta property surface). The 6c variable-mode exhibit needs COLLECTION_ID / MODE_ID / MODES_JSON, which only a variable-collection lookup can supply.
Bound. This is variable collections only — figma_get_variables (or the getLocalVariableCollectionsAsync() equivalent). No component tree walk, no property re-enumeration, no instance inspection.
Call figma_get_variables with format: "summary" to get a lightweight overview of all variable collections in the file. Look for collections whose names contain the component name or common mode-property keywords:
"[ComponentName] shape" — e.g., "Button shape" with modes like Rectangular, Rounded"[ComponentName] density" or "Density" — e.g., "Button density" with modes like Default, Compact, Spacious
For each matching collection, extract:
- Property name: Derive from the collection name (e.g., "Button shape" →
shape, "Density" → density)
- Options: The mode names in the collection (e.g.,
["Rectangular", "Rounded"])
- Default value: The mode named "Default" or "default" if one exists; otherwise the first mode
- Collection name: The full collection name for the annotation note
- Collection ID: The
id field of the collection (e.g., "VariableCollectionId:6028:44006") — needed to apply modes via setExplicitVariableModeForCollection
- Modes: An array of
{ modeId, name } objects for each mode — needed to apply the correct mode per preview instance
Store these as a variableModeProps array alongside variantAxes and booleanProps:
variableModeProps: [
{
name: "shape",
options: ["Rectangular", "Rounded"],
defaultValue: "Rectangular",
collectionName: "Button shape",
collectionId: "VariableCollectionId:1234:5678",
modes: [{ modeId: "1234:0", name: "Rectangular" }, { modeId: "1234:1", name: "Rounded" }]
},
{
name: "density",
options: ["Default", "Compact", "Spacious"],
defaultValue: "Default",
collectionName: "Button density",
collectionId: "VariableCollectionId:6028:44006",
modes: [{ modeId: "6028:0", name: "Default" }, { modeId: "6028:1", name: "Compact" }, { modeId: "6028:2", name: "Spacious" }]
}
]
If no matching collections are found, set variableModeProps to an empty array and proceed.
Step 4c: Child component chapters come from render-meta (discovery walk DELETED)
The legacy Step 4c child-discovery walk is deleted. The old walkForInstances recursion (which walked the default variant for nested INSTANCE children and rebuilt each child's variantAxes / booleanProps / instanceSwapProps from componentPropertyDefinitions) plus its boolean-linkage resolver no longer exist in this skill. Do NOT reintroduce a live tree walk to rediscover child components.
The childComponents array (with controllingBooleanName / controllingBooleanRawKey and the matched controllingBooleanNames) is rebuilt in Step 4 from render-meta.subComponents (constitutive sub-component-SETS only — those with a non-null subCompSetId). Re-read the SCOPE LIMIT in Step 4:
- Only constitutive sub-component-SETS get full child chapters (6e/6f/6g).
- Child surfaces beyond what
render-meta.subComponents carried (non-constitutive nested instances, child booleans/instance-swaps the producer did not capture) degrade gracefully — render what's available and note the gap, pointing to the child's own ./{slug}.md. Never re-extract.
If the rebuilt childComponents array is empty, proceed — there are no constitutive sub-component chapters to exhibit.
Step 4d: Normalize Child Properties (IN MEMORY — no Figma reads)
This is light in-memory reasoning over the render-meta-derived model from Step 4 — no figma_execute, no Figma reads. Apply all four sub-analyses (coupled axes, container-gated booleans, unified slots, sibling booleans) to produce the normalization plan. The algorithm below is the reference logic; run it in memory over the model — it touches no Figma node.
Operate over the variantAxes array from Step 4 as PARENT_AXES, the rebuilt childComponents array from Step 4 as CHILDREN, and the matched controllingBooleanNames array from Step 4 as CONTROLLING_BOOL_NAMES:
const PARENT_AXES = __PARENT_VARIANT_AXES_JSON__;
const CHILDREN = __CHILD_COMPONENTS_JSON__;
const CONTROLLING_BOOL_NAMES = __CONTROLLING_BOOLEAN_NAMES_JSON__;
for (const child of CHILDREN) {
for (const axis of child.variantAxes) {
axis.coupled = false;
for (const pAxis of PARENT_AXES) {
if (axis.name.toLowerCase() === pAxis.name.toLowerCase()) {
const childSet = new Set(axis.options.map(o => o.toLowerCase()));
const parentSet = new Set(pAxis.options.map(o => o.toLowerCase()));
const isSubset = [...childSet].every(o => parentSet.has(o));
if (isSubset) { axis.coupled = true; break; }
}
}
}
}
const unifiedSlotChapters = [];
const unifiedSubBooleanNames = [];
function shortName(boolName, containerName) {
const prefixWords = containerName.toLowerCase().split(/\s+/);
const boolWords = boolName.split(/\s+/);
let stripped = boolWords.filter(w => !prefixWords.includes(w.toLowerCase()));
if (stripped.length === 0) stripped = boolWords;
return stripped.join(' ');
}
function stripVerbs(name) {
return name.replace(/^(Show|Has|With|Enable|Toggle|Display)\s+/i, '');
}
for (const child of CHILDREN) {
if (!child.controllingBooleanName || child.booleanProps.length === 0) continue;
const subBools = child.booleanProps;
const containerBoolName = child.controllingBooleanName;
const containerBoolRawKey = child.controllingBooleanRawKey;
const combos = [];
combos.push({ label: 'None', containerOn: false, subValues: {} });
if (subBools.length <= 5) {
const count = subBools.length;
const total = 1 << count;
const comboEntries = [];
for (let mask = 1; mask < total; mask++) {
const subValues = {};
const onNames = [];
for (let i = 0; i < count; i++) {
const on = Boolean(mask & (1 << i));
subValues[subBools[i].name] = on;
if (on) onNames.push(stripVerbs(shortName(subBools[i].name, containerBoolName)));
}
comboEntries.push({ label: onNames.join(' + '), containerOn: true, subValues, onCount: onNames.length });
}
comboEntries.sort((a, b) => a.onCount - b.onCount);
const capped = comboEntries.length > 5 ? [...comboEntries.slice(0, 4), comboEntries[comboEntries.length - 1]] : comboEntries;
for (const c of capped) { delete c.onCount; combos.push(c); }
} else {
for (const sb of subBools) {
const subValues = {};
for (const s of subBools) subValues[s.name] = (s.name === sb.name);
combos.push({ label: stripVerbs(shortName(sb.name, containerBoolName)), containerOn: true, subValues });
}
const allOn = {};
for (const s of subBools) allOn[s.name] = true;
combos.push({ label: subBools.map(s => stripVerbs(shortName(s.name, containerBoolName))).join(' + '), containerOn: true, subValues: allOn });
}
if (subBools.length === 1) {
combos[1].label = stripVerbs(shortName(subBools[0].name, containerBoolName));
}
const parentBoolDef = CONTROLLING_BOOL_NAMES.includes(containerBoolName);
let defaultLabel = 'None';
if (parentBoolDef) {
const defaultSubValues = {};
for (const sb of subBools) defaultSubValues[sb.name] = sb.defaultValue;
const match = combos.find(c => c.containerOn && Object.entries(c.subValues).every(([k, v]) => defaultSubValues[k] === v));
if (match) defaultLabel = match.label;
}
unifiedSlotChapters.push({
chapterName: child.name + ' -- ' + containerBoolName,
childName: child.name,
containerBoolName,
containerBoolRawKey,
subBooleans: subBools,
previewCombinations: combos,
defaultLabel
});
for (const sb of subBools) unifiedSubBooleanNames.push(sb.name);
}
const siblingBoolChapters = [];
const siblingBoolNames = [];
const consumedByUnified = new Set(unifiedSubBooleanNames);
for (const child of CHILDREN) {
if (child.controllingBooleanName && child.booleanProps.length > 0) continue;
const remaining = child.booleanProps.filter(b => !consumedByUnified.has(b.name));
if (remaining.length < 2) continue;
const combos = [];
const count = remaining.length;
const total = 1 << count;
const comboEntries = [];
for (let mask = 0; mask < total; mask++) {
const subValues = {};
const onNames = [];
for (let i = 0; i < count; i++) {
const on = Boolean(mask & (1 << i));
subValues[remaining[i].name] = on;
if (on) onNames.push(stripVerbs(remaining[i].name));
}
const label = onNames.length === 0 ? 'None' : onNames.join(' + ');
comboEntries.push({ label, subValues, onCount: onNames.length });
}
comboEntries.sort((a, b) => a.onCount - b.onCount);
const capped = comboEntries.length > 6 ? [...comboEntries.slice(0, 4), comboEntries[comboEntries.length - 2], comboEntries[comboEntries.length - 1]] : comboEntries;
for (const c of capped) { delete c.onCount; combos.push(c); }
const defaultSubValues = {};
for (const b of remaining) defaultSubValues[b.name] = b.defaultValue;
const defaultMatch = combos.find(c => Object.entries(c.subValues).every(([k, v]) => defaultSubValues[k] === v));
const defaultLabel = defaultMatch ? defaultMatch.label : 'None';
siblingBoolChapters.push({
chapterName: child.name,
childName: child.name,
booleans: remaining,
previewCombinations: combos,
defaultLabel
});
for (const b of remaining) siblingBoolNames.push(b.name);
}
return {
childComponents: CHILDREN,
unifiedSlotChapters,
unifiedSubBooleanNames,
siblingBoolChapters,
siblingBoolNames
};
Save the normalization output. It produces:
childComponents — Updated with coupled: true flags on child variant axes that mirror parent axes (4d-i). In Step 9 (6e-i), skip axes where coupled === true.unifiedSlotChapters — Array of chapter entries for container + sub-boolean combinations (4d-ii/iii). Each entry has chapterName, childName, containerBoolName, containerBoolRawKey, subBooleans, previewCombinations, and defaultLabel. Rendered in 6f.unifiedSubBooleanNames — Array of sub-boolean names consumed by unified slot chapters. These are skipped in 6e-ii.siblingBoolChapters — Array of chapter entries for sibling boolean combinations (4d-iv). Each entry has chapterName, childName, booleans, previewCombinations, and defaultLabel. Rendered in 6g.siblingBoolNames — Array of boolean names consumed by sibling boolean chapters. These are skipped in 6e-ii.
Label generation rules (handled by the script):
- Sub-boolean short names are derived by stripping the common prefix shared with the container name, plus common verbs ("Show", "Has", "With", "Enable", "Toggle", "Display")
"None" = container off (unified) or all booleans off (sibling)- Multi-on combos are joined with " + "
- Default label is computed from actual boolean default values
Combination cap (handled by the script): Power sets with more than 5-6 entries are capped to the most meaningful combinations (individually-on states, plus the all-on state).
Graceful fallback: If a child has only 1 remaining boolean after filtering (not consumed by unified slots), it is NOT added to siblingBoolChapters — it stays as a standard boolean chapter rendered in 6e-ii.
Step 4e: Validation and Exhibit Planning (IN MEMORY — no Figma reads)
After the model is rebuilt and normalized (Steps 4–4d), perform validation and exhibit planning in memory over the render-meta-derived model — no Figma reads. Follow the Data Validation and Exhibit Planning sections in the instruction file (agent-property-instruction.md).
This step has two phases: Phase A (Data Validation) sanity-checks the render-meta-derived model, Phase B (Exhibit Planning) plans what to render and how. Do NOT rely on visual inspection (Step 10) as the primary safety net — this step is the designated reasoning layer. If a child surface looks incomplete because render-meta.subComponents did not carry it, treat it as the documented SCOPE LIMIT (degrade with a note) — do NOT re-extract to fill it.
Context axis identification — As the first action in Phase B, follow the "Identify context axes" section in the instruction file. Evaluate each variant axis against the heuristics and select 0–1 context axes (rarely 2). Store the result as contextAxis:
contextAxis: { name: "variant", options: ["primary", "subtle"], defaultValue: "primary" }
// or null if no axis qualifies
When contextAxis is non-null:
- The context axis's own exhibit plan entry is
presentation: "illustrate" with template: "6a" (standard, non-contextual). This gives engineers a dedicated reference for the axis options.
- All other
"illustrate" entries use contextual templates (6a-ctx instead of 6a, 6b-ctx instead of 6b).
- Composite chapters use the context rowGroup pattern (see 6a-ctx).
- The
briefDescription should mention the context axis (e.g., "…available in primary and subtle variants").
Produce the exhibitPlan array and contextAxis as documented in the instruction file. Also compose the briefDescription string for the spec header.
After validation and planning, proceed to the pre-render audit.
Step 5: Audit
Re-read the instruction file (agent-property-instruction.md), focusing on:
- Pre-Render Validation Checklist — walk through every item
- Common Mistakes section
- Do NOT section
Check the exhibit plan and corrected data against each rule. Fix any violations before rendering.
Step 6: Navigate to Destination
If the user provided a separate destination file URL:
figma_navigate — Switch to the destination file
If no destination was provided, stay in the current file.
Step 7: Import and Detach Template
If the user provided a cross-file destination URL (navigated in Step 6), run via figma_execute:
const PROPERTY_TEMPLATE_KEY = '__PROPERTY_TEMPLATE_KEY__';
const templateComponent = await figma.importComponentByKeyAsync(PROPERTY_TEMPLATE_KEY);
const instance = templateComponent.createInstance();
const { x, y } = figma.viewport.center;
instance.x = x - instance.width / 2;
instance.y = y - instance.height / 2;
const frame = instance.detachInstance();
frame.name = '__COMPONENT_NAME__ Properties';
figma.currentPage.selection = [frame];
figma.viewport.scrollAndZoomIntoView([frame]);
return { frameId: frame.id };
If no destination was provided (default), run via figma_execute — this places the spec on the component's page, to its right:
const PROPERTY_TEMPLATE_KEY = '__PROPERTY_TEMPLATE_KEY__';
const COMP_NODE_ID = '__COMPONENT_NODE_ID__';
const compNode = await figma.getNodeByIdAsync(COMP_NODE_ID);
let _p = compNode;
while (_p.parent && _p.parent.type !== 'DOCUMENT') _p = _p.parent;
if (_p.type === 'PAGE') await figma.setCurrentPageAsync(_p);
const templateComponent = await figma.importComponentByKeyAsync(PROPERTY_TEMPLATE_KEY);
const instance = templateComponent.createInstance();
const frame = instance.detachInstance();
const GAP = 200;
frame.x = compNode.x + compNode.width + GAP;
frame.y = compNode.y;
frame.name = '__COMPONENT_NAME__ Properties';
figma.currentPage.selection = [frame];
figma.viewport.scrollAndZoomIntoView([frame]);
return { frameId: frame.id, pageId: _p.id, pageName: _p.name };
Replace __COMPONENT_NAME__ with render-meta.component.componentName. Replace __COMPONENT_NODE_ID__ with render-meta.component.compSetNodeId.
Save the returned frameId.
Step 8: Fill Header Fields
Run via figma_execute (replace __FRAME_ID__, __COMPONENT_NAME__, __BRIEF_DESCRIPTION__). Replace __BRIEF_DESCRIPTION__ with the briefDescription composed during Step 4e:
const frame = await figma.getNodeByIdAsync('__FRAME_ID__');
const textNodes = frame.findAll(n => n.type === 'TEXT');
const fontSet = new Set();
const fontsToLoad = [];
for (const tn of textNodes) {
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
}
await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
const compNameFrame = frame.findOne(n => n.name === '#comp-name-anatomy');
if (compNameFrame) {
const t = compNameFrame.findOne(n => n.type === 'TEXT');
if (t) t.characters = '__COMPONENT_NAME__';
}
const descFrame = frame.findOne(n => n.name === '#brief-component-description');
if (descFrame) {
const t = descFrame.findOne(n => n.type === 'TEXT');
if (t) t.characters = '__BRIEF_DESCRIPTION__';
}
const markerExample = frame.findOne(n => n.name === '#marker-example');
if (markerExample) markerExample.visible = false;
return { success: true };
Step 9: Build Property Exhibits
This is the main rendering step. Iterate over the exhibitPlan array produced in Step 4e. Each entry specifies the chapter type, rendering mode, and configuration. Do NOT mechanically iterate variantAxes then booleanProps — the exhibit plan already accounts for matrix chapters, composite chapters, and context axis rendering.
Template routing based on contextAxis:
| Exhibit type | contextAxis is null | contextAxis is non-null |
|---|
| Variant axis chapter | 6a (standard) | 6a-ctx (contextual rows) |
| Boolean chapter | 6b (standard) | 6b-ctx (contextual rows) |
| Composite chapter | 6a (custom OPTIONS) | 6a-ctx (custom OPTIONS + context rows) |
| Sparse matrix | 6a-matrix | 6a-matrix (unchanged) |
| Variable mode | 6c | 6c (unchanged) |
| Child component | 6e/6f/6g | 6e/6f/6g (unchanged) |
When contextAxis is non-null, pass CONTEXT_AXIS_NAME, CONTEXT_OPTIONS, and CONTEXT_DEFAULT to the contextual templates. These values come from the contextAxis object produced in Step 4e.
Run one figma_execute call per exhibit to avoid timeouts. The scripts below are templates — select the appropriate template based on each exhibit entry's template field.
6a: Standard VARIANT axis chapter
For exhibit plan entries with template: "6a" (when contextAxis is null). Also used for composite entries without context — supply a customized OPTIONS array and DEFAULT_PROPS as determined by the exhibit plan:
const FRAME_ID = '__FRAME_ID__';
const COMP_SET_ID = '__COMP_SET_NODE_ID__';
const PROPERTY_NAME = '__PROPERTY_NAME__';
const OPTIONS = __OPTIONS_JSON__;
const DEFAULT_VALUE = '__DEFAULT_VALUE__';
const DEFAULT_PROPS = __DEFAULT_PROPS_JSON__;
const FONT_FAMILY = '__FONT_FAMILY__';
const frame = await figma.getNodeByIdAsync(FRAME_ID);
const chapterTemplate = frame.findOne(n => n.name === '#anatomy-section');
async function loadAllFonts(rootNode) {
const textNodes = rootNode.findAll(n => n.type === 'TEXT');
const fontSet = new Set();
const fontsToLoad = [];
for (const tn of textNodes) {
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
}
await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
}
async function loadFontWithFallback(family, preferredStyle, fallbackStyle) {
fallbackStyle = fallbackStyle || 'Regular';
const allFonts = await figma.listAvailableFontsAsync();
const familyFonts = allFonts.filter(f => f.fontName.family === family);
const match = familyFonts.find(f => f.fontName.style === preferredStyle);
if (match) { await figma.loadFontAsync(match.fontName); return match.fontName; }
const fallback = familyFonts.find(f => f.fontName.style === fallbackStyle);
if (fallback) { await figma.loadFontAsync(fallback.fontName); return fallback.fontName; }
if (familyFonts.length > 0) { await figma.loadFontAsync(familyFonts[0].fontName); return familyFonts[0].fontName; }
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
return { family: 'Inter', style: 'Regular' };
}
const chapter = chapterTemplate.clone();
chapterTemplate.parent.appendChild(chapter);
chapter.name = PROPERTY_NAME;
chapter.visible = true;
try {
await loadAllFonts(chapter);
const sectionName = chapter.findOne(n => n.name === '#section-name');
if (sectionName) {
const t = sectionName.findOne(n => n.type === 'TEXT');
if (t) t.characters = PROPERTY_NAME;
}
const sectionDesc = chapter.findOne(n => n.name === '#optional-section-description');
if (sectionDesc) {
const t = sectionDesc.findOne(n => n.type === 'TEXT');
if (t) t.characters = OPTIONS.length + ' options. Default: ' + DEFAULT_VALUE;
}
const assetPlaceholder = chapter.findOne(n => n.name === '#preview');
while (assetPlaceholder.children.length > 0) {
assetPlaceholder.children[0].remove();
}
assetPlaceholder.layoutWrap = 'WRAP';
assetPlaceholder.counterAxisSpacing = assetPlaceholder.itemSpacing;
const compSet = await figma.getNodeByIdAsync(COMP_SET_ID);
for (const option of OPTIONS) {
const variantProps = {};
for (const [k, v] of Object.entries(DEFAULT_PROPS)) {
variantProps[k] = v;
}
variantProps[PROPERTY_NAME] = option;
let targetVariant = null;
let bestFallback = null;
let bestFallbackScore = -1;
for (const child of compSet.children) {
const vp = child.variantProperties || {};
if (vp[PROPERTY_NAME] !== option) continue;
let score = 0;
let exactMatch = true;
for (const [k, v] of Object.entries(variantProps)) {
if (vp[k] === v) { score++; } else { exactMatch = false; }
}
if (exactMatch) { targetVariant = child; break; }
if (score > bestFallbackScore) { bestFallbackScore = score; bestFallback = child; }
}
if (!targetVariant) targetVariant = bestFallback;
const wrapper = figma.createFrame();
wrapper.name = option;
wrapper.layoutMode = 'VERTICAL';
wrapper.primaryAxisAlignItems = 'CENTER';
wrapper.counterAxisAlignItems = 'CENTER';
wrapper.itemSpacing = 12;
wrapper.fills = [];
wrapper.primaryAxisSizingMode = 'AUTO';
wrapper.counterAxisSizingMode = 'AUTO';
assetPlaceholder.appendChild(wrapper);
if (targetVariant) {
const inst = targetVariant.createInstance();
await loadAllFonts(inst);
wrapper.appendChild(inst);
} else {
const placeholder = figma.createText();
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
placeholder.characters = 'Variant unavailable';
placeholder.fontSize = 12;
placeholder.fills = [{ type: 'SOLID', color: { r: 0.6, g: 0.6, b: 0.6 } }];
wrapper.appendChild(placeholder);
}
const LABEL_FONT = await loadFontWithFallback(FONT_FAMILY, 'Medium');
const label = figma.createText();
label.fontName = LABEL_FONT;
label.characters = option === DEFAULT_VALUE ? option + ' (default)' : option;
label.fontSize = 14;
label.fills = [{ type: 'SOLID', color: { r: 0.29, g: 0.29, b: 0.29 } }];
wrapper.appendChild(label);
}
return { success: true, property: PROPERTY_NAME };
} catch (e) {
chapter.remove();
return { error: e.message, rolledBack: true };
}
6a-ctx: Contextual VARIANT axis chapter
When contextAxis is non-null, use this template instead of 6a for variant chapters. Also used for composite chapters with context. The template adds an outer loop over context axis values, rendering grouped rows inside a vertical container frame. Each row group has a row label and a horizontal instance row.
Replace CONTEXT_AXIS_NAME, CONTEXT_OPTIONS, and CONTEXT_DEFAULT with the context axis data from the exhibit plan. Replace all other placeholders as in 6a:
const FRAME_ID = '__FRAME_ID__';
const COMP_SET_ID = '__COMP_SET_NODE_ID__';
const PROPERTY_NAME = '__PROPERTY_NAME__';
const OPTIONS = __OPTIONS_JSON__;
const DEFAULT_VALUE = '__DEFAULT_VALUE__';
const DEFAULT_PROPS = __DEFAULT_PROPS_JSON__;
const CONTEXT_AXIS_NAME = '__CONTEXT_AXIS_NAME__';
const CONTEXT_OPTIONS = __CONTEXT_OPTIONS_JSON__;
const CONTEXT_DEFAULT = '__CONTEXT_DEFAULT__';
const FONT_FAMILY = '__FONT_FAMILY__';
const frame = await figma.getNodeByIdAsync(FRAME_ID);
const chapterTemplate = frame.findOne(n => n.name === '#anatomy-section');
async function loadAllFonts(rootNode) {
const textNodes = rootNode.findAll(n => n.type === 'TEXT');
const fontSet = new Set();
const fontsToLoad = [];
for (const tn of textNodes) {
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
}
await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
}
async function loadFontWithFallback(family, preferredStyle, fallbackStyle) {
fallbackStyle = fallbackStyle || 'Regular';
const allFonts = await figma.listAvailableFontsAsync();
const familyFonts = allFonts.filter(f => f.fontName.family === family);
const match = familyFonts.find(f => f.fontName.style === preferredStyle);
if (match) { await figma.loadFontAsync(match.fontName); return match.fontName; }
const fallback = familyFonts.find(f => f.fontName.style === fallbackStyle);
if (fallback) { await figma.loadFontAsync(fallback.fontName); return fallback.fontName; }
if (familyFonts.length > 0) { await figma.loadFontAsync(familyFonts[0].fontName); return familyFonts[0].fontName; }
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
return { family: 'Inter', style: 'Regular' };
}
const chapter = chapterTemplate.clone();
chapterTemplate.parent.appendChild(chapter);
chapter.name = PROPERTY_NAME;
chapter.visible = true;
try {
await loadAllFonts(chapter);
const sectionName = chapter.findOne(n => n.name === '#section-name');
if (sectionName) {
const t = sectionName.findOne(n => n.type === 'TEXT');
if (t) t.characters = PROPERTY_NAME;
}
const sectionDesc = chapter.findOne(n => n.name === '#optional-section-description');
if (sectionDesc) {
const t = sectionDesc.findOne(n => n.type === 'TEXT');
if (t) t.characters = OPTIONS.length + ' options across ' + CONTEXT_OPTIONS.length + ' ' + CONTEXT_AXIS_NAME + 's. Default: ' + DEFAULT_VALUE;
}
const assetPlaceholder = chapter.findOne(n => n.name === '#preview');
while (assetPlaceholder.children.length > 0) {
assetPlaceholder.children[0].remove();
}
const compSet = await figma.getNodeByIdAsync(COMP_SET_ID);
const LABEL_FONT = await loadFontWithFallback(FONT_FAMILY, 'Medium');
const ROW_LABEL_FONT = await loadFontWithFallback(FONT_FAMILY, 'Bold');
const contextContainer = figma.createFrame();
contextContainer.name = 'context-groups';
contextContainer.layoutMode = 'VERTICAL';
contextContainer.itemSpacing = 32;
contextContainer.fills = [];
contextContainer.primaryAxisSizingMode = 'AUTO';
contextContainer.counterAxisSizingMode = 'FILL';
assetPlaceholder.appendChild(contextContainer);
for (const ctxValue of CONTEXT_OPTIONS) {
const rowGroup = figma.createFrame();
rowGroup.name = ctxValue;
rowGroup.layoutMode = 'VERTICAL';
rowGroup.itemSpacing = 16;
rowGroup.fills = [];
rowGroup.primaryAxisSizingMode = 'AUTO';
rowGroup.counterAxisSizingMode = 'FILL';
contextContainer.appendChild(rowGroup);
const rowLabel = figma.createText();
rowLabel.fontName = ROW_LABEL_FONT;
rowLabel.characters = ctxValue === CONTEXT_DEFAULT ? ctxValue + ' (default)' : ctxValue;
rowLabel.fontSize = 12;
rowLabel.fills = [{ type: 'SOLID', color: { r: 0.45, g: 0.45, b: 0.45 } }];
rowGroup.appendChild(rowLabel);
const instanceRow = figma.createFrame();
instanceRow.name = ctxValue + '-instances';
instanceRow.layoutMode = 'HORIZONTAL';
instanceRow.layoutWrap = 'WRAP';
instanceRow.itemSpacing = 24;
instanceRow.counterAxisSpacing = 24;
instanceRow.fills = [];
instanceRow.primaryAxisSizingMode = 'AUTO';
instanceRow.counterAxisSizingMode = 'AUTO';
rowGroup.appendChild(instanceRow);
for (const option of OPTIONS) {
const variantProps = { ...DEFAULT_PROPS };
variantProps[PROPERTY_NAME] = option;
variantProps[CONTEXT_AXIS_NAME] = ctxValue;
let targetVariant = null;
let bestFallback = null;
let bestFallbackScore = -1;
for (const child of compSet.children) {
const vp = child.variantProperties || {};
if (vp[PROPERTY_NAME] !== option) continue;
if (vp[CONTEXT_AXIS_NAME] !== ctxValue) continue;
let score = 0;
let exactMatch = true;
for (const [k, v] of Object.entries(variantProps)) {
if (vp[k] === v) { score++; } else { exactMatch = false; }
}
if (exactMatch) { targetVariant = child; break; }
if (score > bestFallbackScore) { bestFallbackScore = score; bestFallback = child; }
}
if (!targetVariant) targetVariant = bestFallback;
const wrapper = figma.createFrame();
wrapper.name = option;
wrapper.layoutMode = 'VERTICAL';
wrapper.primaryAxisAlignItems = 'CENTER';
wrapper.counterAxisAlignItems = 'CENTER';
wrapper.itemSpacing = 12;
wrapper.fills = [];
wrapper.primaryAxisSizingMode = 'AUTO';
wrapper.counterAxisSizingMode = 'AUTO';
instanceRow.appendChild(wrapper);
if (targetVariant) {
const inst = targetVariant.createInstance();
await loadAllFonts(inst);
wrapper.appendChild(inst);
} else {
const placeholder = figma.createText();
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
placeholder.characters = 'N/A';
placeholder.fontSize = 12;
placeholder.fills = [{ type: 'SOLID', color: { r: 0.6, g: 0.6, b: 0.6 } }];
wrapper.appendChild(placeholder);
}
const label = figma.createText();
label.fontName = LABEL_FONT;
label.characters = option === DEFAULT_VALUE ? option + ' (default)' : option;
label.fontSize = 14;
label.fills = [{ type: 'SOLID', color: { r: 0.29, g: 0.29, b: 0.29 } }];
wrapper.appendChild(label);
}
}
return { success: true, property: PROPERTY_NAME };
} catch (e) {
chapter.remove();
return { error: e.message, rolledBack: true };
}
Key differences from 6a:
- Outer loop over
CONTEXT_OPTIONS creates row groups with labels
- A
contextContainer frame inside #preview handles vertical stacking (avoids modifying #preview properties)
variantProps sets both PROPERTY_NAME and CONTEXT_AXIS_NAME for each instance- Variant lookup requires both the property AND context axis to match, with fallback scoring
- Row labels use Bold/12px to distinguish from option labels (Medium/14px)
- N/A placeholders appear when a context × option combination doesn't exist
Composite chapters with context: When a composite chapter (variant axis + related booleans) needs context rendering, use the same 6a-ctx structure. The OPTIONS loop creates instances with custom property combinations (as in the standard composite approach), and the outer CONTEXT_OPTIONS loop wraps everything in context rows. For each composite option, set the variant properties AND the boolean properties on the instance, then also set CONTEXT_AXIS_NAME = ctxValue.
6a-matrix: For a SPARSE VARIANT MATRIX chapter
When the exhibit plan (Step 4e) identified a sparse axis pair, render a matrix chapter plus standalone chapters for both axes. The matrix's primary axis forms the rows, the secondary axis forms the columns. Missing combinations get "N/A" placeholders that occupy the same cell space as real instances for visual alignment. The standalone chapters (6a) give engineers a dedicated reference for each axis in isolation; the matrix shows which cross-product combinations exist.
Grid layout technique: The matrix uses absolute positioning inside a non-auto-layout child frame, nested within the template's #preview frame. This prevents auto-layout from collapsing or misaligning cells when "N/A" placeholders are smaller than real instances.
const FRAME_ID = '__FRAME_ID__';
const COMP_SET_ID = '__COMP_SET_NODE_ID__';
const PRIMARY_AXIS = '__PRIMARY_AXIS_NAME__';
const SECONDARY_AXIS = '__SECONDARY_AXIS_NAME__';
const PRIMARY_OPTIONS = __PRIMARY_OPTIONS_JSON__;
const SECONDARY_OPTIONS = __SECONDARY_OPTIONS_JSON__;
const DEFAULT_PROPS = __DEFAULT_PROPS_JSON__;
const FONT_FAMILY = '__FONT_FAMILY__';
const CHAPTER_NAME = '__CHAPTER_NAME__';
const DESCRIPTION = '__DESCRIPTION__';
const frame = await figma.getNodeByIdAsync(FRAME_ID);
const chapterTemplate = frame.findOne(n => n.name === '#anatomy-section');
async function loadAllFonts(rootNode) {
const textNodes = rootNode.findAll(n => n.type === 'TEXT');
const fontSet = new Set();
const fontsToLoad = [];
for (const tn of textNodes) {
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
}
await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
}
async function loadFontWithFallback(family, preferredStyle, fallbackStyle) {
fallbackStyle = fallbackStyle || 'Regular';
const allFonts = await figma.listAvailableFontsAsync();
const familyFonts = allFonts.filter(f => f.fontName.family === family);
if (familyFonts.length === 0) return { family: 'Inter', style: fallbackStyle };
const pref = familyFonts.find(f => f.fontName.style === preferredStyle);
if (pref) return pref.fontName;
const fb = familyFonts.find(f => f.fontName.style === fallbackStyle);
return fb ? fb.fontName : familyFonts[0].fontName;
}
const chapter = chapterTemplate.clone();
chapter.visible = true;
chapter.name = CHAPTER_NAME;
frame.appendChild(chapter);
try {
await loadAllFonts(chapter);
const titleNode = chapter.findOne(n => n.name === '#property-name' && n.type === 'TEXT');
if (titleNode) titleNode.characters = CHAPTER_NAME;
const descNode = chapter.findOne(n => n.name === '#property-description' && n.type === 'TEXT');
if (descNode) descNode.characters = DESCRIPTION;
const assetPlaceholder = chapter.findOne(n => n.name === '#preview');
while (assetPlaceholder.children.length > 0) assetPlaceholder.children[0].remove();
const compSet = await figma.getNodeByIdAsync(COMP_SET_ID);
const sampleVariant = compSet.children[0];
const sampleInst = sampleVariant.createInstance();
await loadAllFonts(sampleInst);
const CELL_W = Math.ceil(sampleInst.width) + 40;
const CELL_H = Math.ceil(sampleInst.height) + 40;
sampleInst.remove();
const LABEL_H = 20;
const HEADER_H = 24;
const GAP = 8;
const ROW_LABEL_W = 120;
const GRID_LEFT = ROW_LABEL_W + GAP;
const totalCols = SECONDARY_OPTIONS.length;
const totalRows = PRIMARY_OPTIONS.length;
const totalW = GRID_LEFT + totalCols * (CELL_W + GAP);
const totalH = HEADER_H + GAP + totalRows * (CELL_H + LABEL_H + GAP);
assetPlaceholder.layoutWrap = 'WRAP';
const gridFrame = figma.createFrame();
gridFrame.name = CHAPTER_NAME + '-grid';
gridFrame.layoutMode = 'NONE';
gridFrame.fills = [];
gridFrame.resize(totalW, totalH);
assetPlaceholder.appendChild(gridFrame);
const LABEL_FONT = await loadFontWithFallback(FONT_FAMILY, 'Medium');
const HEADER_FONT = await loadFontWithFallback(FONT_FAMILY, 'Bold');
for (let ci = 0; ci < SECONDARY_OPTIONS.length; ci++) {
const header = figma.createText();
header.fontName = HEADER_FONT;
header.characters = SECONDARY_OPTIONS[ci];
header.fontSize = 12;
header.fills = [{ type: 'SOLID', color: { r: 0.4, g: 0.4, b: 0.4 } }];
gridFrame.appendChild(header);
header.x = GRID_LEFT + ci * (CELL_W + GAP) + CELL_W / 2 - header.width / 2;
header.y = 0;
}
for (let ri = 0; ri < PRIMARY_OPTIONS.length; ri++) {
const rowY = HEADER_H + GAP + ri * (CELL_H + LABEL_H + GAP);
const rowLabel = figma.createText();
rowLabel.fontName = LABEL_FONT;
rowLabel.characters = PRIMARY_OPTIONS[ri];
rowLabel.fontSize = 14;
rowLabel.fills = [{ type: 'SOLID', color: { r: 0.29, g: 0.29, b: 0.29 } }];
gridFrame.appendChild(rowLabel);
rowLabel.x = 0;
rowLabel.y = rowY + CELL_H / 2 - rowLabel.height / 2;
for (let ci = 0; ci < SECONDARY_OPTIONS.length; ci++) {
const cellX = GRID_LEFT + ci * (CELL_W + GAP);
const cellY = rowY;
const variantProps = { ...DEFAULT_PROPS };
variantProps[PRIMARY_AXIS] = PRIMARY_OPTIONS[ri];
variantProps[SECONDARY_AXIS] = SECONDARY_OPTIONS[ci];
let targetVariant = null;
for (const child of compSet.children) {
const vp = child.variantProperties || {};
let match = true;
for (const [k, v] of Object.entries(variantProps)) {
if (vp[k] !== v) { match = false; break; }
}
if (match) { targetVariant = child; break; }
}
const wrapper = figma.createFrame();
wrapper.layoutMode = 'VERTICAL';
wrapper.primaryAxisAlignItems = 'CENTER';
wrapper.counterAxisAlignItems = 'CENTER';
wrapper.itemSpacing = 8;
wrapper.fills = [];
wrapper.primaryAxisSizingMode = 'AUTO';
wrapper.counterAxisSizingMode = 'FIXED';
wrapper.resize(CELL_W, CELL_H + LABEL_H);
if (targetVariant) {
const inst = targetVariant.createInstance();
await loadAllFonts(inst);
wrapper.appendChild(inst);
} else {
const naText = figma.createText();
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
naText.characters = 'N/A';
naText.fontSize = 14;
naText.fills = [{ type: 'SOLID', color: { r: 0.7, g: 0.7, b: 0.7 } }];
wrapper.appendChild(naText);
}
gridFrame.appendChild(wrapper);
wrapper.x = cellX;
wrapper.y = cellY;
}
}
return { success: true, chapter: CHAPTER_NAME };
} catch (e) {
chapter.remove();
return { error: e.message, rolledBack: true };
}
N/A placeholder rule: Always render "N/A" text for missing combinations. Never skip the cell or leave it empty — the placeholder preserves the grid's visual scanability and lets the spec consumer immediately see which combinations don't exist.
Cell sizing: Measure a real instance before building the grid. Use the measured dimensions + padding as the fixed cell size. All cells (instance and N/A) use the same width to maintain column alignment.
6b: Standard BOOLEAN property chapter
For exhibit plan entries with template: "6b" (when contextAxis is null).
Skip controlling booleans: Before rendering each parent boolean, check if its name appears in the controllingBooleanNames set built in Step 4 (matched from render-meta.subComponents). If so, skip it — its chapter is produced by 6e as part of the unified child component chapter.
Handle variant-gated booleans: Before rendering, check if the boolean has requiredVariantOverrides (from Step 4a). If so, the base variant for instance creation must match those overrides instead of using the default variant. Replace VARIANT_OVERRIDES with the required overrides object (e.g., {"Behavior": "Interactive"}), or null if the boolean is not variant-gated.
Slot-aware descriptions: Replace __CONTROLS_SLOT_BOOL__ with the boolean's controlsSlot value (true or false). Replace __SLOT_PREFERRED_NAMES_JSON__ with the boolean's slotPreferredNames array (e.g., ["Checkbox", "Radio"]), or [] if empty. When a boolean controls a SLOT, the description reads "Controls slot: {name} (accepts: {preferred})" instead of "Controls layer: {name}".
For each remaining boolean property, run via figma_execute:
const FRAME_ID = '__FRAME_ID__';
const COMP_SET_ID = '__COMP_SET_NODE_ID__';
const PROPERTY_NAME = '__PROPERTY_NAME__';
const DEFAULT_VALUE = __DEFAULT_BOOL_VALUE__;
const ASSOCIATED_LAYER = '__ASSOCIATED_LAYER__';
const CONTROLS_SLOT = __CONTROLS_SLOT_BOOL__;
const SLOT_PREFERRED_NAMES = __SLOT_PREFERRED_NAMES_JSON__;
const VARIANT_OVERRIDES = __VARIANT_OVERRIDES_OR_NULL__;
const FONT_FAMILY = '__FONT_FAMILY__';
async function loadAllFonts(rootNode) {
const textNodes = rootNode.findAll(n => n.type === 'TEXT');
const fontSet = new Set();
const fontsToLoad = [];
for (const tn of textNodes) {
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
}
await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
}
async function loadFontWithFallback(family, preferredStyle, fallbackStyle) {
fallbackStyle = fallbackStyle || 'Regular';
const allFonts = await figma.listAvailableFontsAsync();
const familyFonts = allFonts.filter(f => f.fontName.family === family);
const match = familyFonts.find(f => f.fontName.style === preferredStyle);
if (match) { await figma.loadFontAsync(match.fontName); return match.fontName; }
const fallback = familyFonts.find(f => f.fontName.style === fallbackStyle);
if (fallback) { await figma.loadFontAsync(fallback.fontName); return fallback.fontName; }
if (familyFonts.length > 0) { await figma.loadFontAsync(familyFonts[0].fontName); return familyFonts[0].fontName; }
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
return { family: 'Inter', style: 'Regular' };
}
const frame = await figma.getNodeByIdAsync(FRAME_ID);
const chapterTemplate = frame.findOne(n => n.name === '#anatomy-section');
const chapter = chapterTemplate.clone();
chapterTemplate.parent.appendChild(chapter);
chapter.name = PROPERTY_NAME;
chapter.visible = true;
try {
await loadAllFonts(chapter);
const sectionName = chapter.findOne(n => n.name === '#section-name');
if (sectionName) {
const t = sectionName.findOne(n => n.type === 'TEXT');
if (t) t.characters = PROPERTY_NAME;
}
const sectionDesc = chapter.findOne(n => n.name === '#optional-section-description');
if (sectionDesc) {
const t = sectionDesc.findOne(n => n.type === 'TEXT');
const defaultStr = DEFAULT_VALUE ? 'true' : 'false';
let layerStr = '';
if (CONTROLS_SLOT) {
layerStr = '. Controls slot: ' + ASSOCIATED_LAYER;
if (SLOT_PREFERRED_NAMES.length > 0) layerStr += ' (accepts: ' + SLOT_PREFERRED_NAMES.join(', ') + ')';
} else if (ASSOCIATED_LAYER) {
layerStr = '. Controls layer: ' + ASSOCIATED_LAYER;
}
const gateStr = VARIANT_OVERRIDES ? '. Requires ' + Object.entries(VARIANT_OVERRIDES).map(([k,v]) => k + ' = ' + v).join(', ') : '';
if (t) t.characters = 'Boolean toggle. Default: ' + defaultStr + layerStr + gateStr;
}
const assetPlaceholder = chapter.findOne(n => n.name === '#preview');
while (assetPlaceholder.children.length > 0) {
assetPlaceholder.children[0].remove();
}
assetPlaceholder.layoutWrap = 'WRAP';
assetPlaceholder.counterAxisSpacing = assetPlaceholder.itemSpacing;
const compNode = await figma.getNodeByIdAsync(COMP_SET_ID);
let baseVariant;
if (VARIANT_OVERRIDES && compNode.type === 'COMPONENT_SET') {
const defaultVProps = (compNode.defaultVariant || compNode.children[0]).variantProperties || {};
const targetProps = { ...defaultVProps, ...VARIANT_OVERRIDES };
baseVariant = null;
let bestScore = -1;
for (const child of compNode.children) {
const vp = child.variantProperties || {};
let score = 0;
let exact = true;
for (const [k, v] of Object.entries(targetProps)) {
if (vp[k] === v) { score++; } else { exact = false; }
}
if (exact) { baseVariant = child; break; }
if (score > bestScore) { bestScore = score; baseVariant = child; }
}
} else {
baseVariant = compNode.type === 'COMPONENT_SET'
? (compNode.defaultVariant || compNode.children[0])
: compNode;
}
const LABEL_FONT = await loadFontWithFallback(FONT_FAMILY, 'Medium');
for (const boolVal of [true, false]) {
const wrapper = figma.createFrame();
wrapper.name = PROPERTY_NAME + ' = ' + boolVal;
wrapper.layoutMode = 'VERTICAL';
wrapper.primaryAxisAlignItems = 'CENTER';
wrapper.counterAxisAlignItems = 'CENTER';
wrapper.itemSpacing = 12;
wrapper.fills = [];
wrapper.primaryAxisSizingMode = 'AUTO';
wrapper.counterAxisSizingMode = 'AUTO';
assetPlaceholder.appendChild(wrapper);
const inst = baseVariant.createInstance();
await loadAllFonts(inst);
wrapper.appendChild(inst);
for (const [rawKey, val] of Object.entries(inst.componentProperties)) {
const cleanKey = rawKey.split('#')[0];
if (cleanKey === PROPERTY_NAME) {
inst.setProperties({ [rawKey]: boolVal });
await loadAllFonts(inst);
break;
}
}
const label = figma.createText();
label.fontName = LABEL_FONT;
const isDefault = boolVal === DEFAULT_VALUE;
label.characters = String(boolVal) + (isDefault ? ' (default)' : '');
label.fontSize = 14;
label.fills = [{ type: 'SOLID', color: { r: 0.29, g: 0.29, b: 0.29 } }];
wrapper.appendChild(label);
}
return { success: true, property: PROPERTY_NAME };
} catch (e) {
chapter.remove();
return { error: e.message, rolledBack: true };
}
6b-ctx: Contextual BOOLEAN property chapter
When contextAxis is non-null, use this template instead of 6b for boolean chapters. It wraps the true/false toggle in context rows so the developer sees how the boolean looks across all context values.
Skip controlling booleans and handle variant-gated booleans using the same rules as 6b.
For each remaining boolean property, run via figma_execute:
const FRAME_ID = '__FRAME_ID__';
const COMP_SET_ID = '__COMP_SET_NODE_ID__';
const PROPERTY_NAME = '__PROPERTY_NAME__';
const DEFAULT_VALUE = __DEFAULT_BOOL_VALUE__;
const ASSOCIATED_LAYER = '__ASSOCIATED_LAYER__';
const CONTROLS_SLOT = __CONTROLS_SLOT_BOOL__;
const SLOT_PREFERRED_NAMES = __SLOT_PREFERRED_NAMES_JSON__;
const VARIANT_OVERRIDES = __VARIANT_OVERRIDES_OR_NULL__;
const CONTEXT_AXIS_NAME = '__CONTEXT_AXIS_NAME__';
const CONTEXT_OPTIONS = __CONTEXT_OPTIONS_JSON__;
const CONTEXT_DEFAULT = '__CONTEXT_DEFAULT__';
const FONT_FAMILY = '__FONT_FAMILY__';
async function loadAllFonts(rootNode) {
const textNodes = rootNode.findAll(n => n.type === 'TEXT');
const fontSet = new Set();
const fontsToLoad = [];
for (const tn of textNodes) {
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
}
await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
}
async function loadFontWithFallback(family, preferredStyle, fallbackStyle) {
fallbackStyle = fallbackStyle || 'Regular';
const allFonts = await figma.listAvailableFontsAsync();
const familyFonts = allFonts.filter(f => f.fontName.family === family);
const match = familyFonts.find(f => f.fontName.style === preferredStyle);
if (match) { await figma.loadFontAsync(match.fontName); return match.fontName; }
const fallback = familyFonts.find(f => f.fontName.style === fallbackStyle);
if (fallback) { await figma.loadFontAsync(fallback.fontName); return fallback.fontName; }
if (familyFonts.length > 0) { await figma.loadFontAsync(familyFonts[0].fontName); return familyFonts[0].fontName; }
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
return { family: 'Inter', style: 'Regular' };
}
const frame = await figma.getNodeByIdAsync(FRAME_ID);
const chapterTemplate = frame.findOne(n => n.name === '#anatomy-section');
const chapter = chapterTemplate.clone();
chapterTemplate.parent.appendChild(chapter);
chapter.name = PROPERTY_NAME;
chapter.visible = true;
try {
await loadAllFonts(chapter);
const sectionName = chapter.findOne(n => n.name === '#section-name');
if (sectionName) {
const t = sectionName.findOne(n => n.type === 'TEXT');
if (t) t.characters = PROPERTY_NAME;
}
const sectionDesc = chapter.findOne(n => n.name === '#optional-section-description');
if (sectionDesc) {
const t = sectionDesc.findOne(n => n.type === 'TEXT');
const defaultStr = DEFAULT_VALUE ? 'true' : 'false';
let layerStr = '';
if (CONTROLS_SLOT) {
layerStr = '. Controls slot: ' + ASSOCIATED_LAYER;
if (SLOT_PREFERRED_NAMES.length > 0) layerStr += ' (accepts: ' + SLOT_PREFERRED_NAMES.join(', ') + ')';
} else if (ASSOCIATED_LAYER) {
layerStr = '. Controls layer: ' + ASSOCIATED_LAYER;
}
const gateStr = VARIANT_OVERRIDES ? '. Requires ' + Object.entries(VARIANT_OVERRIDES).map(([k,v]) => k + ' = ' + v).join(', ') : '';
if (t) t.characters = 'Boolean toggle across ' + CONTEXT_OPTIONS.length + ' ' + CONTEXT_AXIS_NAME + 's. Default: ' + defaultStr + layerStr + gateStr;
}
const assetPlaceholder = chapter.findOne(n => n.name === '#preview');
while (assetPlaceholder.children.length > 0) {
assetPlaceholder.children[0].remove();
}
const compNode = await figma.getNodeByIdAsync(COMP_SET_ID);
const LABEL_FONT = await loadFontWithFallback(FONT_FAMILY, 'Medium');
const ROW_LABEL_FONT = await loadFontWithFallback(FONT_FAMILY, 'Bold');
const contextContainer = figma.createFrame();
contextContainer.name = 'context-groups';
contextContainer.layoutMode = 'VERTICAL';
contextContainer.itemSpacing = 32;
contextContainer.fills = [];
contextContainer.primaryAxisSizingMode = 'AUTO';
contextContainer.counterAxisSizingMode = 'FILL';
assetPlaceholder.appendChild(contextContainer);
for (const ctxValue of CONTEXT_OPTIONS) {
const rowGroup = figma.createFrame();
rowGroup.name = ctxValue;
rowGroup.layoutMode = 'VERTICAL';
rowGroup.itemSpacing = 16;
rowGroup.fills = [];
rowGroup.primaryAxisSizingMode = 'AUTO';
rowGroup.counterAxisSizingMode = 'FILL';
contextContainer.appendChild(rowGroup);
const rowLabel = figma.createText();
rowLabel.fontName = ROW_LABEL_FONT;
rowLabel.characters = ctxValue === CONTEXT_DEFAULT ? ctxValue + ' (default)' : ctxValue;
rowLabel.fontSize = 12;
rowLabel.fills = [{ type: 'SOLID', color: { r: 0.45, g: 0.45, b: 0.45 } }];
rowGroup.appendChild(rowLabel);
const instanceRow = figma.createFrame();
instanceRow.name = ctxValue + '-instances';
instanceRow.layoutMode = 'HORIZONTAL';
instanceRow.layoutWrap = 'WRAP';
instanceRow.itemSpacing = 24;
instanceRow.counterAxisSpacing = 24;
instanceRow.fills = [];
instanceRow.primaryAxisSizingMode = 'AUTO';
instanceRow.counterAxisSizingMode = 'AUTO';
rowGroup.appendChild(instanceRow);
const defaultVProps = (compNode.defaultVariant || compNode.children[0]).variantProperties || {};
const baseProps = VARIANT_OVERRIDES ? { ...defaultVProps, ...VARIANT_OVERRIDES } : { ...defaultVProps };
baseProps[CONTEXT_AXIS_NAME] = ctxValue;
let baseVariant = null;
let bestScore = -1;
for (const child of compNode.children) {
const vp = child.variantProperties || {};
let score = 0;
let exact = true;
for (const [k, v] of Object.entries(baseProps)) {
if (vp[k] === v) { score++; } else { exact = false; }
}
if (exact) { baseVariant = child; break; }
if (score > bestScore) { bestScore = score; baseVariant = child; }
}
if (!baseVariant) {
const skipLabel = figma.createText();
skipLabel.fontName = LABEL_FONT;
skipLabel.characters = 'Not available for ' + ctxValue;
skipLabel.fontSize = 12;
skipLabel.fills = [{ type: 'SOLID', color: { r: 0.6, g: 0.6, b: 0.6 } }];
instanceRow.appendChild(skipLabel);
continue;
}
for (const boolVal of [true, false]) {
const wrapper = figma.createFrame();
wrapper.name = PROPERTY_NAME + ' = ' + boolVal;
wrapper.layoutMode = 'VERTICAL';
wrapper.primaryAxisAlignItems = 'CENTER';
wrapper.counterAxisAlignItems = 'CENTER';
wrapper.itemSpacing = 12;
wrapper.fills = [];
wrapper.primaryAxisSizingMode = 'AUTO';
wrapper.counterAxisSizingMode = 'AUTO';
instanceRow.appendChild(wrapper);
const inst = baseVariant.createInstance();
await loadAllFonts(inst);
wrapper.appendChild(inst);
for (const [rawKey, val] of Object.entries(inst.componentProperties)) {
const cleanKey = rawKey.split('#')[0];
if (cleanKey === PROPERTY_NAME) {
inst.setProperties({ [rawKey]: boolVal });
await loadAllFonts(inst);
break;
}
}
const label = figma.createText();
label.fontName = LABEL_FONT;
const isDefault = boolVal === DEFAULT_VALUE;
label.characters = String(boolVal) + (isDefault ? ' (default)' : '');
label.fontSize = 14;
label.fills = [{ type: 'SOLID', color: { r: 0.29, g: 0.29, b: 0.29 } }];
wrapper.appendChild(label);
}
}
return { success: true, property: PROPERTY_NAME };
} catch (e) {
chapter.remove();
return { error: e.message, rolledBack: true };
}
Key differences from 6b:
- Outer loop over
CONTEXT_OPTIONS creates row groups with labels
- Base variant lookup includes
CONTEXT_AXIS_NAME = ctxValue in the target props
- When no base variant exists for a context value (sparse), the row shows "Not available for {ctxValue}" instead of failing
- Same
contextContainer → rowGroup → instanceRow nesting as 6a-ctx
6c: For each VARIABLE MODE property
If variableModeProps is not empty, render a visual chapter for each. Variable mode properties are controlled via Figma variable modes at the container level. To produce visual previews, create a wrapper frame for each mode option, place a component instance inside, and call wrapper.setExplicitVariableModeForCollection(collection, modeId) on the wrapper so the instance inherits the mode.
Important — collection object, not string ID: The Figma plugin API in incremental mode requires the actual collection object for setExplicitVariableModeForCollection, not a string ID. The script below fetches the collection object via getLocalVariableCollectionsAsync().
Important — clearing baked-in modes: Some components have explicit variable modes set directly on their root or internal nodes. Instances created from such components inherit these baked-in modes, which override the wrapper's mode. After creating each instance, the script recursively clears explicit modes for the target collection so the instance defers to the wrapper.
For each variable mode property, run via figma_execute:
const FRAME_ID = '__FRAME_ID__';
const COMP_SET_ID = '__COMP_SET_NODE_ID__';
const PROPERTY_NAME = '__PROPERTY_NAME__';
const DEFAULT_VALUE = '__DEFAULT_VALUE__';
const COLLECTION_NAME = '__COLLECTION_NAME__';
const COLLECTION_ID = '__COLLECTION_ID__';
const MODES = __MODES_JSON__;
const FONT_FAMILY = '__FONT_FAMILY__';
async function loadAllFonts(rootNode) {
const textNodes = rootNode.findAll(n => n.type === 'TEXT');
const fontSet = new Set();
const fontsToLoad = [];
for (const tn of textNodes) {
try {
const fn = tn.fontName;
if (fn && fn !== figma.mixed && fn.family) {
const key = fn.family + '|' + fn.style;
if (!fontSet.has(key)) { fontSet.add(key); fontsToLoad.push(fn); }
}
} catch {}
}
await Promise.all(fontsToLoad.map(f => figma.loadFontAsync(f).catch(() => {})));
}
async function loadFontWithFallback(family, preferredStyle, fallbackStyle) {
fallbackStyle = fallbackStyle || 'Regular';
const allFonts = await figma.listAvailableFontsAsync();
const familyFonts = allFonts.filter(f => f.fontName.family === family);
const match = familyFonts.find(f => f.fontName.style === preferredStyle);
if (match) { await figma.loadFontAsync(match.fontName); return match.fontName; }
const fallback = familyFonts.find(f => f.fontName.style === fallbackStyle);
if (fallback) { await figma.loadFontAsync(fallback.fontName); return fallback.fontName; }
if (familyFonts.length > 0) { await figma.loadFontAsync(familyFonts[0].fontName); return familyFonts[0].fontName; }
await figma.loadFontAsync({ family: 'Inter', style: 'Regular' });
return { family: 'Inter', style: 'Regular' };
}
const frame = await figma.getNodeByIdAsync(FRAME_ID);
const chapterTemplate = frame.findOne(n => n.name === '#anatomy-section');
const chapter = chapterTemplate.clone();
chapterTemplate.parent.appendChild(chapter);
chapter.name = PROPERTY_NAME;
chapter.visible = true;
try {
const collections = await figma.variables.getLocalVariableCollectionsAsync();
const collection = collections.find(c => c.id === COLLECTION_ID);
if (!collection) {
chapter.remove();
return { error: 'Variable collection not found: ' + COLLECTION_ID };
}
function clearModesRecursive(node, col) {
try { node.clearExplicitVariableModeForCollection(col); } catch {}
if ('children' in node) {
for (const child of node.children) clearModesRecursive(child, col);
}
}
await loadAllFonts(chapter);
const sectionName = chapter.findOne(n => n.name === '#section-name');
if (sectionName) {
const t = sectionName.findOne(n => n.type === 'TEXT');
if (t) t.characters = PROPERTY_NAME;
}
