MCP tool calls fail in queue mode: `Tool node "<name>" does not have supplyData method` + `Worker tool execution timeout`

Bug Description Summary When using MCP Server Trigger in queue mode (EXECUTIONS_MODE=queue), MCP tool calls intermittently fail with: • Tool node "<name>" does not have supplyData method • mcp_tool_execution_error • Worker tool execution timeout This breaks MCP requests with protocol errors (e.g. 424). Environment • n8n image: docker.n8n.io/n8nio/n8n:stable (currently 2.9.2) • runners image: n8nio/runners:stable (currently 2.9.2) • Deployment: Docker Compose • Topology: main web + worker + redis + postgres + runners • EXECUTIONS_MODE=queue • N8N_RUNNERS_MODE=external Error examples What I found In queue mode, MCP execution path is forced to queued strategy in: /usr/local/lib/node_modules/n8n/dist/scaling/scaling.service.js Then tool execution is delegated to worker path (job-processor.js), where it can throw: Temporary workaround (works) Disable queued MCP strategy (let MCP use direct strategy in web…