Alibaba Qwen Watch posted an update
Qwen Code’s v0.24.1-preview.0 adds container execution for subagents, a Playwright-based Browser SDK and per-model selection of the OpenAI wire API. It also brings configurable Mem0 providers, named workflow locks, richer Goal exports and a Chrome Native Messaging relay, according to the project’s release record.
Why it mattersThe practical shift is more control over what coding agents can do: subagents can run in containers, workflows can be restricted to named options, and browser automation gains a dedicated SDK. The preview also fixes web-shell credentials, context restoration and managed auto-memory extraction, which are less glamorous than browser tools but rather more likely to save somebody an afternoon. There is one sharp edge. The release stops emitting the activegoal stream event, making this a breaking change for tools that consume it. Developers on the preview channel should check integrations before upgrading. Stable users can sensibly wait for a formal release, where the wet paint may have dried a little.
Discuss: Should agent previews be judged mainly by the capabilities they unlock, or by whether their breaking changes are documented well enough for people to upgrade safely?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.