yunfengzhou-hub opened a new issue, #1258: URL: https://github.com/apache/flink-agents/issues/1258
### Search before asking - [x] I searched in the [issues](https://github.com/apache/flink-agents/issues) and found nothing similar. ### Description Sub-agent support has been built out under this umbrella, but the documentation still only explains how to declare one. `docs/content/docs/development/multi-agent/subagents.md` covers registration and the two ways a call can start, and stops there — a user cannot work out from it how a call behaves once it is running, so the capability is not usable from the docs alone. Document what a user has to know to run a sub-agent in a job: - How a call ends: what comes back, how the caller learns the outcome, and what happens if an action finishes with a call still outstanding. - How a call behaves across a failover: whether the work repeats, and what a result must satisfy to be replayed rather than re-run. - How to write an external sub-agent, and how to choose between the two execution models the page currently collapses into "the appropriate runtime base class". - How to report failure: what a sub-agent should return versus raise, what the delegating model sees when a call fails, and how a result is converted on its way back. - What a child agent can see: its memory relative to the caller's, and whether it may have sub-agents of its own. - How a reader gets here from the pages that mention sub-agents, and back. Every statement must match the current implementation, and every example must actually run. ### Are you willing to submit a PR? - [x] I'm willing to submit a PR! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
