Add AI prompt tool for natural language graph operations

New MCP tool and portal UI for executing natural language instructions
against the memory graph via Ollama (qwen3-coder:30b). Single LLM call
generates a JSON action plan which is executed sequentially.

Supports 8 action types: add_node, update_node, remove_node, add_edge,
remove_edge, bulk_tag, reorganize, query. Actions can reference previous
results via $result[N].field interpolation. Uses /api/chat with few-shot
assistant example, format:json, and temperature:0 for reliable output.

src/mcp/index.ts

@@ -367,6 +367,21 @@ server.tool(
   }
 );
 
+// --- memory_prompt ---
+import { interpretAndExecute } from '../core/prompt/interpreter';
+
+server.tool(
+  'memory_prompt',
+  'Execute a natural language instruction against the memory graph. Uses AI to generate and run an action plan. Returns log of actions performed.',
+  {
+    prompt: z.string().describe('Natural language instruction'),
+  },
+  async ({ prompt }) => {
+    const result = await interpretAndExecute(prompt);
+    return { content: [{ type: 'text' as const, text: serialize(result) }] };
+  }
+);
+
 async function main() {
   const transport = new StdioServerTransport();
   await server.connect(transport);