closing-linear-looplisted

# Closing Linear Loop ## Overview This is the shared-skill replacement for the legacy Claude Code `/linear:done` workflow. Use it when the user is intentionally declaring that an issue is complete and wants the loop advanced correctly. ## Read First Before acting, read: - `.claude/skills/linear-loop/SKILL.md` - `.claude/skills/linear-loop/config.json` These define completion routing, loop continuity, project transitions, and the monitor follow-up rules. ## Process 1. Identify the issue to close: - use the explicitly provided issue ID when present - otherwise infer only from strong local context - ask a short bounded question if the target issue is still ambiguous 2. Build a completion summary: - what changed - relevant files or spec links when applicable - any follow-up work or monitor creation 3. Add the completion comment to the issue. 4. Move the issue to Done and update agent-state labels per the loop conventions. 5. Create follow-up monitor work when required by the issue type. 6. Read the issue’s `## On Completion` section when present. 7. Run a project pulse check using the Loop Continuity rules from `linear-loop/SKILL.md`. 8. Clean up the workspace when applicable: - if the work ran in a dedicated git worktree (check the spec's `04-implementation.md`, or the current directory) and its branch is merged, offer `/worktree:remove <branch> --delete-branch` - if the session is currently inside that worktree (`git rev-parse --git-dir --git-co