calm-artist-7206
09/09/2026, 6:25 AMfresh-machine-37269
09/09/2026, 6:25 AMfresh-machine-37269
09/09/2026, 6:53 AMall-hands-bot) wasn't approving because at that point (1) the PR was still a draft and the HUMAN: description section was an unexpanded placeholder, and (2) there were open review threads flagging store=True silently no-oping and the previous_response_id continuation misreporting the model. Those were process/correctness blockers, not "merge it" verdicts ā so the bot left a COMMENT (with "Worth merging" text) rather than an approval.
What's changed since then (all now resolved):
⢠store=True now returns a clear 400 instead of silently no-oping.
⢠The unsafe previous_response_id continuation approximation was removed entirely, which also eliminated the model-misreporting concern.
⢠Both inline review threads are marked resolved.
⢠The HUMAN: section now has the author's own words, and the PR is no longer a draft.
My check: CI is green (statusCheckRollup SUCCESS, including agent-server/cross/stress/openapi suites), the stateless boundary and explicit rejection of unsupported options look sound, and there's strong live official-SDK + concurrency + restart/replay coverage. No major issues found.
I've submitted an APPROVE review on the current head commit: github.com/OpenHands/software-agent-sdk/pull/4694#ā¦
One note: the merge button may still show "blocked" pending a maintainer authorizing the workflow ā the required unresolved-review-threads check and review requirement now look satisfied, so it should clear on the next evaluation.calm-artist-7206
09/09/2026, 7:11 AMfresh-machine-37269
09/09/2026, 7:40 AM