Actions2
- Tool Actions
Tool → Call Tool
AI-generatedSummary
Invoke a specific MCP tool by name with JSON-formatted arguments, sending the request to the configured HackNotice MCP server and returning its result, optionally splitting the output content into individual n8n items.
Inputs
- Tool Name or ID (required) — Select or specify the name or ID of the MCP tool to call. The list dynamically loads from the connected MCP server.
- Arguments (JSON) (required) — Provide a JSON object matching the selected tool's input schema, representing arguments to pass to the tool. Use empty object '{}' if there are no arguments.
- Output Each Content Item Separately — Choose whether to emit one workflow item per element of the MCP tool's result
contentarray (true), or to emit a single item containing the entire result (false). Defaults to true. - Fail on MCP Tool Error — Enable to cause the node to throw an error if the MCP tool's response signals an error (
isError=true), triggering error workflows. Defaults to true.
Output shape
one or more n8n items representing the MCP tool invocation result
If 'Output Each Content Item Separately' is enabled and the result contains a 'content' array, each entry is emitted as a separate item containing that entry augmented with metadata fields indicating the tool name and error status. Otherwise, a single item is emitted containing the full tool result with fields 'toolName', 'isError', 'content', and optionally 'structuredContent'. If the MCP server response has 'isError=true' and 'Fail on MCP Tool Error' is enabled, the node throws an error. Errors or unexpected input cause the node to throw or emit error details depending on the node's error handling settings.