Resolving Merge Conflicts

1. Inspect the operation state. Check Git status, the merge or rebase todo state, the commit ranges on both sides, and every file Git reports as conflicted.
2. Trace each side's intent. For every conflict, inspect the commits that introduced both sides and locate the originating PRs, issues, or tickets from links, commit messages, or repository history.
3. Resolve every hunk within the established scope. Determine the intended resolution from available project artifacts, including issues, pull requests, specifications, ADRs, documentation, and commit history. Make ordinary implementation decisions needed to carry out that intent, preserving both intended behaviors when they can coexist. If a resolution would exceed the established scope or the intended behavior cannot be determined, preserve completed resolutions and unrelated work, leave the merge or rebase paused, and report the blocker and remaining conflicts. Do not reset, abort, or discard work merely because a later hunk is blocked.
4. Run automated checks. Use the repository-defined commands and order; when no order is defined, typically run typecheck, then tests, then formatting. Fix anything the merge or rebase broke.
5. Finish the operation. Stage only resolved conflict paths and intentional fixes required by the merge or rebase; preserve unrelated staged, unstaged, and untracked work. Inspect the complete staged candidate; stop if continuing would commit unrelated work. Run git merge --continue or git rebase --continue as applicable; if Git requires a direct merge commit, create it. For a rebase, repeat these steps for each subsequent conflict until Git reports that all commits have been rebased.