genable-mcp 0.1.1 → 0.1.3

package/dist/tools-schema.json

@@ -183,7 +183,7 @@
   },
   {
     "name": "replace_props",
-    "description": "Batch search-and-replace property values in a subtree.\n\nExamples:\n  replace_props({node: \"1:2\", rules: [{prop: \"fillColor\", from: \"#FFF\", to: \"#000\"}]})\n  replace_props({node: \"1:2\", rules: [\n    {prop: \"fillColor\", from: \"#FFF\", to: \"#000\"},\n    {prop: \"fontSize\", from: \"14\", to: \"16\"}\n  ]})",
+    "description": "Bulk find-and-replace property values across a subtree (target node + all descendants). Destructive batch mutation — no preview, no undo across many nodes. Returns per-rule match counts.\n\nUse when:\n- Theming pass: change every #FFF fill to #000 across a screen\n- Token migration: bump every fontSize from 14 to 16\n- Normalizing values left inconsistent by earlier passes\n- The alternative is N targeted single-node calls (set_text / set_fill / edit)\n\nReturns: { data: { replacements: [{ rule: 0, matched: 12 }, { rule: 1, matched: 0 }] } }\n\nParameters beyond schema:\n- `node` is the subtree root; search recurses into all descendants (depth-first).\n- Each rule's `from` is an EXACT-match string (no substring, no regex). For typed props (fontSize, opacity), pass values as strings — the executor coerces.\n- Zero matches do NOT error — they return matched: 0. Sanity-check with discover_props first if you're unsure values exist.\n\nSkip when:\n- Updating a single known node — use set_text / set_fill / set_stroke / set_layout for type-aware single-intent edits, or edit for generic.\n- Values are variable-bound (tokens) — replace_props bypasses bindings; use bind_variable to swap the token instead.\n- You need partial / fuzzy match — replace_props is exact-only; you'll need find_nodes + a loop.\n\nExamples:\n  // single rule, white -> black\n  replace_props({node: \"1:2\", rules: [{prop: \"fillColor\", from: \"#FFF\", to: \"#000\"}]})\n\n  // batch theme update — both rules applied in one pass\n  replace_props({node: \"1:2\", rules: [\n    {prop: \"fillColor\", from: \"#FFF\", to: \"#000\"},\n    {prop: \"fontSize\", from: \"14\", to: \"16\"}\n  ]})",
     "parameters": {
       "type": "object",
       "properties": {
@@ -730,7 +730,7 @@
   },
   {
     "name": "create_instance",
-    "description": "Create an instance of a component.\n\nExamples:\n  create_instance({node: \"1:2\"})\n  create_instance({node: \"1:2\", parent: \"1:4\"})",
+    "description": "Create an instance of an existing component. Mutates the canvas — appends a new InstanceNode as the last child of `parent` (or the active page root if omitted). The instance is LINKED to the component master, so future component edits propagate. Returns the new instance's nodeId.\n\nUse when:\n- Spawning runtime copies of a Component master (buttons, list items, cards)\n- Reusing a design-system component in a fresh layout\n- Programmatic instantiation outside a jsx() tree-build\n\nReturns: { data: { id: \"5:42\", name: \"Button\", componentId: \"1:2\" } }\n\nParameters beyond schema:\n- `node` must be a Component node (not Frame, Text, or another Instance). Discover IDs with find_nodes({ type: \"COMPONENT\" }).\n- `parent` optional. If parent is auto-layout, the instance enters the flow and inherits sizing rules. If omitted, the instance is placed at the active page root with detached position — may overlap existing content; set explicit position with edit afterwards.\n\nSkip when:\n- Duplicating a non-component node — instance creation will fail; use clone_node instead.\n- Building a subtree from scratch — use jsx with <instance ref=\"ComponentName\"/> for atomic single-call construction.\n\nExamples:\n  create_instance({node: \"1:2\"})                    // at page root\n  create_instance({node: \"1:2\", parent: \"1:4\"})     // inside frame 1:4",
     "parameters": {
       "type": "object",
       "properties": {
@@ -1203,7 +1203,7 @@
   },
   {
     "name": "session_note",
-    "description": "Read / write your own session scratchpad. Persists across turns within one design session (\"New Design\" resets it). Notes are how state carries from this turn into the next — the next session loads them as context.\n\nActions:\n  action: \"read\"   → read({key}) returns the current value (empty string if unset)\n  action: \"write\"  → write({key, value}) replaces or deletes (pass value:\"\" to delete)\n  action: \"list\"   → list({}) returns [{key, chars}] for all existing notes\n\nSlots — FORWARD-LOOKING (what we plan / commit to):\n  - plan       — this turn's intent + step outline (write at turn start)\n  - decisions  — locked choices: style picked + reason, accent token, font scale, hero treatment, etc. (write BEFORE jsx)\n  - brand      — durable brand notes pulled from a project design.md (if user supplied one)\n  - todo       — TRULY unfinished work for the next turn (omit if everything shipped)\n\nSlots — BACKWARD-LOOKING (what happened, write AT TURN END):\n  - failures   — tool calls that failed this turn + how you worked around them.\n                 Example: \"jsx items='stretch' rejected (DSL valid: center|start|end|space-between|baseline); retried with 'center'.\"\n                 If a failure repeats a class you've seen before, name the class.\n  - gotchas    — validator warnings you noticed but chose not to fix + why.\n                 Example: \"4 LOW_CONTRAST on nav links (2.5:1) — deliberate for ambient-grey style; revisit if user complains.\"\n                 Also: magic numbers / hand-tuned positions and what motivated them.\n                 Example: \"Glow ellipses at (-180, 220) / (1060, 70) — placed half outside frame to bleed in.\"\n  - learnings  — surprises about this codebase / DSL / Figma API.\n                 Example: \"radial-gradient(circle at X% Y%) rejected — DSL only takes the simple form. Same trap as CSS-prior bleed elsewhere.\"\n\nREQUIRED behavior:\n  - First turn: write at least `decisions` BEFORE any jsx (commit-before-act).\n  - Every turn end: write at least ONE of `failures / gotchas / learnings` if ANY of these happened this turn:\n      • a tool call returned an error\n      • a tool call returned warnings you chose not to fix\n      • you hand-tuned a coordinate / size / color away from a value the model would have picked\n      • you found a DSL behavior that surprised you\n    \"All clean, no carry-over\" is almost never accurate — at least one backward slot belongs.\n  - jsx that uses a color / font / size should round-trip through `decisions` (token traceability).\n\nExamples:\n  session_note({action: \"write\", key: \"decisions\", value: \"Style: fintech-dark.\\nAccent: #3B82F6 (style.accent — NOT indigo).\\nFont: Space Grotesk display / Inter body.\\nHero treatment: split (overrides anatomy VERTICAL — desktop convention).\\nH1 size: 32 (style.display).\"})\n  session_note({action: \"write\", key: \"failures\", value: \"jsx #6 align='stretch' rejected (DSL valid: center|start|end|space-between|baseline) — retried with 'center'. Same CSS-prior class as gradient: model has CSS values that DSL doesn't accept.\"})\n  session_note({action: \"write\", key: \"gotchas\", value: \"8 LOW_CONTRAST warnings on nav links + secondary CTA (2.5:1 against dark bg). Left as-is — user prompt didn't require WCAG pass; flagging in case next iteration tightens contrast.\"})\n  session_note({action: \"write\", key: \"learnings\", value: \"layoutPositioning='absolute' children render correctly inside auto-layout frame, but parent needs clipsContent=true to suppress overflow on big glow ellipses.\"})\n  session_note({action: \"read\", key: \"decisions\"})\n  session_note({action: \"list\"})",
+    "description": "Read / write your own session scratchpad. Persists across turns within one design session (\"New Design\" resets it). Notes are how state carries from this turn into the next — the next session loads them as context.\n\nActions:\n  action: \"read\"   → read({key}) returns the current value (empty string if unset)\n  action: \"write\"  → write({key, value}) replaces or deletes (pass value:\"\" to delete)\n  action: \"list\"   → list({}) returns [{key, chars}] for all existing notes\n\nSlots — FORWARD-LOOKING (what we plan / commit to):\n  - plan       — this turn's intent + step outline (write at turn start)\n  - decisions  — locked choices: style picked + reason, accent token, font scale, hero treatment, etc. (write BEFORE jsx)\n  - brand      — durable brand notes pulled from a project design.md (if user supplied one)\n  - todo       — TRULY unfinished work for the next turn (omit if everything shipped)\n\nSlots — BACKWARD-LOOKING (what happened, write AT TURN END — AUTO-MERGE on write):\n  - failures   — tool calls that failed this turn + how you worked around them.\n                 Example: \"jsx items='stretch' rejected (DSL valid: center|start|end|space-between|baseline); retried with 'center'.\"\n                 If a failure repeats a class you've seen before, name the class.\n  - gotchas    — validator warnings you noticed but chose not to fix + why.\n                 Example: \"4 LOW_CONTRAST on nav links (2.5:1) — deliberate for ambient-grey style; revisit if user complains.\"\n                 Also: magic numbers / hand-tuned positions and what motivated them.\n                 Example: \"Glow ellipses at (-180, 220) / (1060, 70) — placed half outside frame to bleed in.\"\n  - learnings  — surprises about this codebase / DSL / Figma API.\n                 Example: \"radial-gradient(circle at X% Y%) rejected — DSL only takes the simple form. Same trap as CSS-prior bleed elsewhere.\"\n\nBACKWARD-LOOKING slots auto-merge: writing to `failures / gotchas / learnings` appends to prior content after a \"---\" divider — your new value never silently overwrites accumulated retrospective notes. To replace fresh, first `write({value: \"\"})` to clear, then write again. The result `data.merged=true` flag confirms when merge happened.\n\nREQUIRED behavior:\n  - First turn: write at least `decisions` BEFORE any jsx (commit-before-act).\n  - Subsequent turns: `read` `decisions` and `learnings` BEFORE any jsx — those slots survive across turns and your conversationHistory won't carry the full content reliably. Your turn's adjacent <system-reminder> snapshot also surfaces current notes, but explicit `read` proves you considered them.\n  - Every turn end: write at least ONE of `failures / gotchas / learnings` if ANY of these happened this turn:\n      • a tool call returned an error\n      • a tool call returned warnings you chose not to fix\n      • you hand-tuned a coordinate / size / color away from a value the model would have picked\n      • you found a DSL behavior that surprised you\n    \"All clean, no carry-over\" is almost never accurate — at least one backward slot belongs.\n  - jsx that uses a color / font / size should round-trip through `decisions` (token traceability).\n\nExamples:\n  session_note({action: \"write\", key: \"decisions\", value: \"Style: fintech-dark.\\nAccent: #3B82F6 (style.accent — NOT indigo).\\nFont: Space Grotesk display / Inter body.\\nHero treatment: split (overrides anatomy VERTICAL — desktop convention).\\nH1 size: 32 (style.display).\"})\n  session_note({action: \"write\", key: \"failures\", value: \"jsx #6 align='stretch' rejected (DSL valid: center|start|end|space-between|baseline) — retried with 'center'. Same CSS-prior class as gradient: model has CSS values that DSL doesn't accept.\"})\n  session_note({action: \"write\", key: \"gotchas\", value: \"8 LOW_CONTRAST warnings on nav links + secondary CTA (2.5:1 against dark bg). Left as-is — user prompt didn't require WCAG pass; flagging in case next iteration tightens contrast.\"})\n  session_note({action: \"write\", key: \"learnings\", value: \"layoutPositioning='absolute' children render correctly inside auto-layout frame, but parent needs clipsContent=true to suppress overflow on big glow ellipses.\"})\n  session_note({action: \"read\", key: \"decisions\"})\n  session_note({action: \"list\"})",
     "parameters": {
       "type": "object",
       "properties": {

package/package.json

@@ -1,11 +1,11 @@
 {
   "name": "genable-mcp",
-  "version": "0.1.1",
-  "mcpName": "io.github.muse40007/genable-mcp",
+  "version": "0.1.3",
+  "mcpName": "io.github.musepy/genable-mcp",
   "description": "Write-side MCP server for Figma — build, edit, restructure, and search Figma designs from Claude Code, Cursor, Cline, or any MCP client. Complements Figma's official MCP (which is read-only). 41 tools covering tree creation (JSX), variables/tokens, components, cross-page navigation, and visual verification.",
-  "author": "musec <muse40007@gmail.com>",
+  "author": "musepy <muse40007@gmail.com> (https://github.com/musepy)",
   "bugs": {
-    "url": "https://github.com/muse40007/figma-ai-generator-dogfood/issues"
+    "url": "https://github.com/musepy/genable/issues"
   },
   "main": "dist/index.js",
   "bin": {
@@ -65,10 +65,10 @@
   "publishConfig": {
     "access": "public"
   },
-  "homepage": "https://github.com/muse40007/figma-ai-generator-dogfood#readme",
+  "homepage": "https://github.com/musepy/genable/tree/main/tools/mcp-server#readme",
   "repository": {
     "type": "git",
-    "url": "git+https://github.com/muse40007/figma-ai-generator-dogfood.git",
+    "url": "git+https://github.com/musepy/genable.git",
     "directory": "tools/mcp-server"
   },
   "license": "MIT"