coordinate-worktreeslisted

# Coordinate Worktrees Land independent work without losing bytes or silently choosing between competing product intents. A clean integration branch is not proof that every worktree's progress remains reachable. ## Safety - Freeze or explicitly coordinate active writers before integration. A status snapshot is immediately stale while another agent continues editing. - Never use broad reset, clean, checkout-overwrite, blanket stash, forced branch switch, rebase of shared work, or `ours`/`theirs` conflict strategy. - Do not delete worktrees, branches, stashes, backup refs, or temporary evidence without explicit authorization after reachability is proven. - Treat untracked, ignored-but-important, symlink, mode, submodule, sparse-checkout, detached-HEAD, and sequencer state as first-class data. - Use `organized-commits` for dirty lanes when available. Preserve unrelated user changes and stop when intent is ambiguous. ## Workflow 1. Identify the integration target from remote/default-branch evidence and fetch current refs without changing worktrees. 2. Inventory every worktree: path, branch/HEAD, upstream, ahead/behind, merge-base, staged/unstaged/untracked state, ignored-but-important files, stash references, and merge/rebase/cherry-pick state. 3. Classify each lane as clean committed work, dirty-only work, already reachable, patch-equivalent duplicate, divergent, or active/blocked. Use patch-ID plus history and semantic inspection; subjects alone do not prove duplication. 4