GitHub took the Copilot cloud agent integration for Linear out of preview and made it generally available, which means teams that manage engineering work in Linear can now assign an issue directly to an autonomous Copilot agent instead of a human, without leaving Linear's interface. Once assigned, the agent reads the issue, works in its own ephemeral environment powered by GitHub Actions, and opens a draft pull request, streaming progress updates back into the Linear issue's activity timeline the whole time so the assigning engineer can watch what it's doing without needing to tab over to GitHub. When the work is done, it requests a PR review like any other contributor would. For teams already using Linear as their source of truth for planning, this closes a workflow gap that previously required manually copying issues into GitHub, kicking off an agent there, and copying status back, friction that made just delegate this ticket to an agent a multi-step chore rather than a one-click action. The GA release also adds real configurability that the preview lacked: teams can pick which underlying model handles a given delegated issue, use repository-defined custom agents rather than a generic default, and set base and working branches directly from Linear, with workspace- and team-level settings letting engineering leads standardize how autonomous work gets handled across a whole organization. That last part matters more than it sounds, as async coding agents become a normal part of ticket triage, having centralized policy over which repos, branches, and models agents are allowed to touch is what turns a cool demo into something a security-conscious engineering org will actually turn on for real backlogs.