@modern-fireman-99755 I have some PRs enabling tool_preset for evaluation. This could be useful to your work on evaluating agents with the delegation tool, since you can create a tool preset for delegation. I’ll assign you as a reviewer: • https://github.com/OpenHands/evaluation/pull/259 • https://github.com/OpenHands/benchmarks/pull/419 • https://github.com/OpenHands/software-agent-sdk/pull/2077 I already tested on this SWE-bench run for gemini, and it seems to have worked, so I think it’s probably functional: https://openhands-ai.slack.com/archives/C09QGUDQVTL/p1771228149045599