Hotfix
A small, urgent production fix released outside the normal schedule because waiting would cause more harm.
What it is
A hotfix is a small, urgent change released to repair a serious production problem outside the normal release schedule. It starts from the exact code currently running, changes as little as possible, passes an abbreviated but real set of checks, and ships as a patch release. The urgency changes the queue, not the definition of done. Vincent Driessen's 2010 Gitflow model is where the named hotfix branch comes from: cut it from the production tag, then merge it back into both the release line and the development line.
Reach for it when users are blocked, data is at risk, or a security flaw cannot wait for the next planned release. If restoring the previous build is safer, roll back first and prepare the hotfix with production healthy rather than debugging under outage pressure.
Gotcha: a hotfix branch easily becomes a secret fork of production. Merge or cherry-pick the fix back into the normal development line immediately, tag what actually shipped, and never patch files directly on a server. Skipping every check often turns one incident into two.
Ask AI for it
Prepare a hotfix from the Git release tag that is currently in production. Create a hotfix branch from that exact tag, make only the smallest safe correction, add a regression test, and run the existing GitHub Actions checks. Bump the SemVer patch number, create an annotated release tag, deploy its immutable artifact, and run the production smoke tests. Then cherry-pick the fix onto main and record the incident, tag, commit SHA, deploy time, and rollback command in the release notes.