Dispatch worktree leaves untracked agentic_os.egg-info, uv.lock, .claude - settle gitignore / uv.lock policy #79
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Problem
The dispatch worktree
.dispatch-worktrees/agentic-os/issue-149ended in a state thatcoily dispatch reaprefuses to clean - three untracked paths:Branch is already merged into
origin/main. The work shipped. The untracked dirt is what blocks cleanup.This is a sibling failure mode to
coilysiren/coily#122(which covers the.claude/lockdown-deny.sh+.claude/settings.jsonmodified-files pattern). This one is untracked-files, and the file set is agentic-os-specific.Fix
Three independent calls, all in this repo:
agentic_os.egg-info/- Python build artifact, should be in.gitignore. Standard*.egg-info/entry. Almost certainly just an oversight.uv.lock- decide the policy. uv encourages committing lock files; if that's the call here, commit it and add a CI check that regeneration is a no-op. If not, gitignore it. Either way, "untracked and surprising" is the failure mode to remove..claude/- upstream fix is coilysiren/coily#122 (dispatch should not touch.claude/*in worktrees at all). Once that lands, the worktree-side.claude/won't materialize. Until then, adding.claude/to this repo's.gitignoreis the bandaid - flagged because the long-term answer lives in coily, not here.Out of scope
agentic-os/issue-149worktree. Safe togit worktree remove --forceonce (1) and (2) are decided so the next dispatch doesn't reintroduce the same dirty state.Context
Surfaced 2026-05-27 while inspecting eight stale dispatch worktrees. Six fit the coily#122 pattern. This one and
gauntlet/issue-39are the outliers. Gauntlet's a fully orphaned dir (no.gitlink, separate cleanup, not worth filing).