Agno v2.2.10 adds run_context as the standard state-sharing parameter across Workflows and introduces yield_run_output for Agent/Team runs.
$ git clone --branch v2.2.10 https://github.com/agno-agi/agno.git # already have the repo? check out this version: $ git checkout v2.2.10
run_context into a custom Python step to share and mutate state across a Workflow without relying on session-level globals.from agno.workflow import Workflow, RunContext
def my_step(run_context: RunContext) -> str:
run_context.state['processed'] = True
return 'done'
wf = Workflow(steps=[my_step])
wf.run() yield_run_output flag instead of the deprecated yield_run_response.for event in agent.run('Summarise this report', yield_run_output=True, stream=True):
print(event) - ›Adds
yield_run_outputflag on Agent/Teamrun/arunfunctions as the replacement for the to-be-deprecatedyield_run_response. - ›Promotes
run_contextas the recommended parameter for reading and modifying state across all Workflow surfaces — Agents, Teams, tools, steps, and custom Python functions used in steps. - ›Improves event streaming from custom executor steps inside Workflows, including better handling of Agent/Team events emitted by a custom executor.