repo-reset

Codex form of the /repo:reset command. Other commands appear as $repo-<verb> skills. Helper scripts and .claude/ paths below are relative to the repository root and exist when the Claude Code surface is installed alongside this one.

$repo-reset — Back to Baseline

The end-of-task ritual: review and prune stale git state, sync with the remote, and land back on the default branch with a clean working tree. The reversible steps (fetch, land on the default branch) run by default; nothing irreversible — dropping a stash, deleting a branch or worktree — ever happens without an explicit opt-in and the permanent-loss check.

Usage

$repo-reset                    # Run the reversible baseline steps; keep all stashes, land on default
$repo-reset --ask              # Confirm each step before acting
$repo-reset --prune            # Also delete confirmed-safe branches/worktrees (after the loss check)

Two halves

The steps below split cleanly in two, and naming the split matters because callers can schedule the halves separately:

• Sync-and-switch (reversible — steps 1 and 4): the working-tree safety check, git fetch --all --prune, and landing on an up-to-date default branch. Nothing is removed.
• Pruning (gated — steps 2 and 3): stash review and branch/worktree review. These can permanently remove work, so they keep the explicit opt-in and the [[branches]] permanent-loss check.

Run standalone, $repo-reset always runs all four steps in order, exactly as documented — this split changes nothing here. It exists because [[all]] may run the sync-and-switch half early (before its Docs stage) when the current branch is fully pushed and behind the default branch, so that its doc stages don't edit a stale checkout; the pruning half still runs last there, with its gates intact.

Steps

1. Working tree safety check