GitHub Expands Copilot Context and Model Controls in Slack and Teams
The public-preview integrations can use richer conversation context, check for similar issues and preserve a selected model across a thread.
Edited by Tyronne Panaino
GitHub updated its Copilot integrations for Slack and Microsoft Teams on September 25, adding more conversation context, model controls and links between a discussion and the GitHub work created from it. The public preview is available to organizations on Copilot Business and Enterprise plans, although GitHub says some capabilities are still rolling out and may not be present in every workspace.
The change matters to engineering teams that begin work in a collaboration channel but need the resulting issue or agent task to remain traceable. It also gives users more control over the model used in a conversation and gives Slack administrators a way to steer work toward default owners and repositories.
Channel context can travel into GitHub work
In Slack, Copilot can now use supported files, attachments and message links as context. In Microsoft Teams, the integration can use inline images, forwarded-message context, and channel and thread history. GitHub says Copilot also checks for similar issues before creating a new one.
When the integration does create GitHub work, it can include a direct link to the result and preserve a link back to the original conversation. That two-way trail should make it easier for a developer reading the issue to recover the discussion that led to it. The announcement establishes the linking behaviour, but it does not measure how reliably the system identifies a truly duplicate issue or extracts the right context from a long thread.
Model and repository choices become more explicit
Users can switch the model for the next message, and Copilot keeps that choice for the rest of the conversation. Slack users can also set default owners and repositories. These controls separate two decisions that were easy to blur in a chat-driven workflow: which model should handle the task, and where the resulting work belongs.
Model choice is still a workflow setting rather than proof of better output. The fetched evidence does not compare model quality, latency or error rates inside Slack or Teams. Teams should validate the selected model against the task and review the repository and owner before an issue or implementation plan becomes operational work.
Long-running work gets clearer recovery signals
GitHub says the update improves implementation-plan status, handling of interrupted or stale replies, and reconnection after an idle conversation. Teams receives additional fixes for retained thread history, duplicate answers, converted images, user-owned repositories and large channels. Slack gets recovery, repository-picker and code-channel fixes, plus safer repository switching so an old session cannot continue acting in a superseded repository.
Those changes address continuity and stale-state risks, but they remain vendor-reported product behaviour. The announcement does not provide independent reliability data or a guarantee that every interrupted task resumes with complete context.
Availability and limits
The preview is limited to organizations on GitHub Copilot Business and Enterprise. Usage draws from existing Copilot entitlements and can be managed through existing cloud-agent budgets. Slack setup requires an enabled Copilot cloud-agent policy and the GitHub app for Slack; Teams additionally requires cloud sandboxes. Gradual rollout means administrators should confirm the controls in their own workspace before documenting them as universally available.
Status
Confirmed. GitHub documents the new context, traceability, model and recovery controls. Internal confidence is medium because availability and performance are supported by one first-party announcement and were not independently tested in this run.
Sources
Update note: Last reviewed 2026-09-26. We will revise this post if GitHub changes plan access, rollout scope or the documented context and recovery controls.
Sources
Drafted with AI assistance from source briefs; reviewed for citation completeness and label accuracy.