Content by Abi Komma, Charles Kim, Vivek Bhadauria, Morteza Ziyadi, Kavitha Budumuri, Vinutha Karanth, Lori Li, Shivank Goel, Yingying Zhao, Yi Fang (1)

Abi Komma and coauthors explain how Microsoft Foundry Skill Optimization turns production traces into reviewable agent skills without fine-tuning. They describe its public preview, evaluation pipeline, and internal benchmark results, including where generated rubrics help, where they can mislead, and why developers retain control over deployment.
Community

End of content

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please reload the page.