MCP-Creator-MCP

Instructions

Save a creation workflow for reuse.

Args:
    name: Workflow name
    description: Workflow description
    steps: list of workflow steps

Returns:
    Confirmation message

Implementation Reference

• main.py:248-279 (handler)

  MCP tool handler for 'save_workflow'. Registers the tool and implements the logic by delegating to the workflow_engine instance from the request context.

  @mcp.tool()
  async def save_workflow(
      ctx: Context,
      name: str,
      description: str,
      steps: list[dict],
  ) -> str:
      """
      Save a creation workflow for reuse.
  
      Args:
          name: Workflow name
          description: Workflow description
          steps: list of workflow steps
  
      Returns:
          Confirmation message
      """
      try:
          workflow_engine = ctx.request_context.lifespan_context["workflow_engine"]
  
          workflow_id = await workflow_engine.save_workflow(
              name=name,
              description=description,
              steps=steps,
          )
  
          return f"✅ Workflow '{name}' saved successfully (ID: {workflow_id})"
  
      except Exception as e:
          logger.error(f"Failed to save workflow: {e}")
          return f"❌ Error saving workflow: {str(e)}"

• src/mcp_creator/workflows/workflow_engine.py:82-105 (helper)

  Core helper method in WorkflowEngine class that implements the actual workflow saving logic, including ID generation, object creation, in-memory storage, and persistence to disk.

  async def save_workflow(
      self,
      name: str,
      description: str,
      steps: list[dict[str, Any]],
  ) -> str:
      """Save workflow with automatic ID generation."""
      workflow_id = str(uuid4())[:8]
  
      # Convert step dictionaries to WorkflowStep objects
      workflow_steps = [WorkflowStep(**step) for step in steps]
  
      workflow = Workflow(
          name=name,
          description=description,
          steps=workflow_steps,
      )
  
      # Save to memory and disk
      self.workflows[workflow_id] = workflow
      await self._persist_workflow(workflow_id, workflow)
  
      logger.info(f"Saved workflow: {name} ({workflow_id})")
      return workflow_id

• main.py:248-248 (registration)

  The @mcp.tool() decorator registers the save_workflow function as an MCP tool.

  @mcp.tool()