purushah commented on PR #964: URL: https://github.com/apache/flink-agents/pull/964#issuecomment-5384652531
Thanks @wenjin272 and @weiqingy for the careful reviews and the merge — the feature came out much stronger. A heads-up on what's next: the `Strategies.llm(...)` follow-up we agreed to defer in #897 — framework-managed LLM-as-judge, where the engine runs the judge call on its own durable/metered/observable chat path — has been in the works alongside this review cycle for a while now, and it's nearly ready. The shape is exactly what the #897 thread settled on: the judge executes under a durable `"judge:<router>"` call with engine retries and token attribution, verdicts are constrained to candidate names, judge failures honor the error-handling strategy, and a mistyped judge model fails at plan construction rather than at runtime. I'll open the PR soon. -- 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]
