Sub-workflow ignores updated queries from parent agent on subsequent calls

Below is some guidance to help resolve the issue where the sub‑workflow is reusing the first query rather than picking up each new query sent by the agent.

---

Check for Pinned Example Data in the Sub‑Workflow Trigger

If you set up your sub‑workflow with the “Load data into sub‑workflow before building” feature, you may have pinned example data in the Execute Workflow Trigger node. While this lets you build with real data, it can cause the trigger node to default to the pinned data rather than any new incoming query.

Solution:
• Open your sub‑workflow.
• Select the Execute Workflow Trigger node.
• Remove or unpin any example data so that the node always uses the actual data passed from the parent workflow.

---

Use the Correct Execution Mode

On the parent workflow side, the Execute Workflow node can be configured either as “Run once with all items” or “Run once for each item.” If you intend to run the sub‑workflow multiple times with different queries, setting it to “Run once for each item” ensures that every input is treated separately.

Tip:
• Edit your Execute Workflow node in the parent workflow and verify that the Mode parameter is set to “Run once for each item.”

---

Ensure That Settings Are in Sync

It’s important that:

• The sub‑workflow is free of errors (if there are errors, it might abort and behave unexpectedly).
• The parent workflow is passing the new query data as input to the Execute Workflow node during each call.

By confirming these settings, you help n8n properly pass along the new query data on every execution rather than reusing the previously pinned or first query.

---

With these adjustments, your sub‑workflow should correctly receive and process the query data sent by your agent on each call. If you continue to have issues, you might also verify that no caching is occurring on your agent side and that each request is being sent cleanly to the parent workflow.