@incident to make the change, and it opens a pull request for you to review — built on everything the investigation already knows.
Asking for a change
In the incident channel, tag@incident and describe the change you want — fix a bug it identified, revert a risky change, add a guard. It opens a pull request against the right repository and posts its progress back into the channel as it works; when the PR is up, it’s linked to the incident like any other resource.
Because the request runs with the full investigation behind it, you don’t have to re-explain the problem. The agent making the change starts from the investigation’s root-cause finding, the evidence behind it, and the code the investigation already read — so the change targets the actual cause rather than guessing from a one-line instruction. That means you can be brief:
@incident open a pull request to fix this.Or point it at something specific when you already know the fix:
@incident this is the missing timeout on the payments client — add a sensible one and open a PR.You can keep iterating in the channel: ask for adjustments and the same pull request is updated.
@incident can you also add a test that covers the timeout?
Making code changes needs a connected code repository with write access. See Code
repositories.
Who makes the change
By default, incident.io makes the change itself: a coding agent runs in an isolated, sandboxed container, makes the edit, and opens the pull request. This is the right default if your team doesn’t already have its own coding agent — it needs no setup beyond connecting your repositories.Delegated agents
If your team already uses an external coding agent, you can delegate the implementation to it instead. incident.io still owns everything around the change — the investigation, the conversation in the channel, the progress updates, and linking the finished pull request back to the incident — and hands off only the step of writing the code. The point of delegating is that the change is made with all the setup you’ve already built in that tool: your rules and conventions, the MCP servers you’ve connected (including incident.io’s own, so the agent can pull live incident context), and your development environment. The fix comes from the agent your team has already tuned to your codebase, rather than a generic one.Cursor
Connect Cursor with an API key, and code change requests are handed to a Cursor cloud agent that makes the change and opens the pull request. It runs with your Cursor configuration — including any MCP servers you’ve set up for your repositories — and the resulting pull request is linked back to the incident exactly as the default flow would.Cursor is the delegated agent available today. If you’d like to use another — such as Devin — get in
touch and we’ll work with you.
Automatic code changes
Investigations can go one step further and propose a fix without being asked. When an investigation reaches high confidence in a root cause that’s addressable in code, it can open a pull request itself and announce it in the channel — turning a diagnosis straight into a proposed fix for you to review. This is off by default. If you’d like investigations to propose code changes automatically, get in touch and we’ll enable it for your organization.Related
Code repositories
Connect GitHub or GitLab, and choose read or read-and-write access.
Incident channel experience
Everything else you can ask
@incident during an incident.