Yes —

{{ flow_run }}

in an automation parameter is typically rendered as a string, not passed through as a native dict/object. If you want Flow B to receive structured data, the supported pattern is to wrap the Jinja output in Prefect's JSON hydration format:

{
  "__prefect_kind": "json",
  "value": {
    "__prefect_kind": "jinja",
    "template": "{{ flow_run | tojson }}"
  }
}

That does two things: 1.

{{ flow_run | tojson }}

serializes the triggering flow run to JSON 2.

__prefect_kind: "json"

tells Prefect to parse that JSON back into a structured parameter So in your automation's

RunDeployment

action, you'd do something like:

parameters={
  "upstream_flow_run": {
    "__prefect_kind": "json",
    "value": {
      "__prefect_kind": "jinja",
      "template": "{{ flow_run | tojson }}"
    }
  }
}

Then Flow B can accept:

@flow
def flow_b(upstream_flow_run: dict):
    print(upstream_flow_run["id"])

Recommended approach: instead of passing the entire

flow_run

object, pass only the fields you actually need:

parameters={
  "trigger_info": {
    "__prefect_kind": "json",
    "value": {
      "__prefect_kind": "jinja",
      "template": "{{ {'id': flow_run.id, 'name': flow_run.name, 'parameters': flow_run.parameters} | tojson }}"
    }
  }
}

That tends to be more stable and easier to work with. A few notes: - Simple scalar values can use plain Jinja:

"name": {
    "__prefect_kind": "jinja",
    "template": "{{ flow_run.name }}"
  }

- For dicts/lists/bools/numbers where you want native types preserved, use the

json

+

jinja

+

tojson

pattern -

flow_run

here refers to the run that triggered the automation If you want, I can also show you a complete automation example for either: - Python SDK automations, or - deployment trigger configuration in YAML/UI.