Tools
cp__call_workflows
Execute multiple workflows in parallel with the cp__call_workflows builtin
The cp__call_workflows builtin dispatches several workflows concurrently, enforcing the configured concurrency ceiling and emitting detailed per-workflow results. Use it for regional fan-out, tenant processing, or staged orchestration.
Input fields
| Field | Type | Description |
|---|---|---|
workflows (required) | array | Ordered list of workflow configurations mirroring cp__call_workflow. Each entry supports workflow_id, optional input, optional initial_task_id, and optional timeout_ms. |
Usage example
{
"workflows": [
{
"workflow_id": "process.region",
"input": { "region": "us-east-1" }
},
{
"workflow_id": "process.region",
"input": { "region": "eu-west-1" },
"timeout_ms": 180000
}
]
}The builtin limits concurrency to runtime.native_tools.call_workflows.max_concurrent. Adjust the configuration to tune throughput.
Remediation hints
- Provide at least one entry in the
workflowsarray. - Ensure every entry includes a valid
workflow_idand satisfies schema expectations. - Use per-workflow
timeout_msvalues for long-running regions instead of relying on the default. - Inspect
results[n].errorfor remediation hints when a workflow fails.