docs(pr-discussions): archive PR 3307

docs/pr-discussions/PR-3307-docs-rules-add-worktree-force-remove-guard-to-claim-acquire.md

@@ -0,0 +1,93 @@
+---
+pr_number: 3307
+title: "docs(rules): add worktree force-remove guard to claim-acquire rule"
+author: "AceHack"
+state: "MERGED"
+created_at: "2026-05-14T23:53:55Z"
+merged_at: "2026-05-14T23:55:43Z"
+closed_at: "2026-05-14T23:55:43Z"
+head_ref: "otto-claim-rule-worktree-force-remove-guard-2026-05-14"
+base_ref: "main"
+archived_at: "2026-05-14T23:58:53Z"
+archive_tool: "tools/pr-preservation/archive-pr.ts"
+---
+
+# PR #3307: docs(rules): add worktree force-remove guard to claim-acquire rule
+
+## PR description
+
+## Summary
+
+Extends [.claude/rules/claim-acquire-before-worktree-work.md](/.claude/rules/claim-acquire-before-worktree-work.md) with a new \"Worktree force-remove guard\" section covering an empirical failure mode that the existing rule didn't anticipate.
+
+## Empirical anchor
+
+[docs/hygiene-history/ticks/2026/05/14/1813Z.md](/docs/hygiene-history/ticks/2026/05/14/1813Z.md) (Otto-CLI's tick shard) documents:
+
+- Otto-CLI tried to checkout Otto-Desktop's PR #3153 branch to investigate a Codex thread
+- Got \`fatal: already used by worktree at /private/tmp/zeta-otto-id-alloc\`
+- Force-removed the worktree to take over
+- Found Otto-Desktop had already authored the fix (\`1636908\`); just resolved the thread
+
+The legitimate intent (review-thread resolution) was covered by the existing DOES-NOT-APPLY clause. The mechanism (force-remove) wasn't. This PR adds the missing guard.
+
+## Three operational alternatives
+
+| Approach | When | Cost |
+|---|---|---|
+| 1. New worktree at distinct path | Default — almost always works | ~30s worktree create + 4400 file checkout |
+| 2. \`gh api\` / GraphQL for branch-state ops | Thread resolution, comment posting, PR metadata | Zero — no checkout needed |
+| 3. Bus-mediated worktree handoff | Rare must-checkout cases | Coordination cost; bus advisory envelope |
+
+## Composition
+
+- Composes with the existing \"When this rule applies\" / \"DOES NOT APPLY\" framing
+- Tightens the rule without invalidating any prior application
+- Empirically grounded — not speculative
+
+## Test plan
+
+- [ ] CI passes (markdownlint + memory frontmatter checks)
+- [ ] Future Otto cold-boots read the new section at session start
+- [ ] Next cross-Otto branch-coordination event exercises the discipline
+
+🤖 Generated with [Claude Code](https://claude.com/claude-code)
+
+## Reviews
+
+### COMMENTED — @copilot-pull-request-reviewer (2026-05-14T23:57:44Z)
+
+## Pull request overview
+
+Extends the claim-acquire rule with guidance for avoiding destructive takeover of another worktree when Git reports a branch is already in use.
+
+**Changes:**
+- Adds a “Worktree force-remove guard” section.
+- Recommends alternate worktree/API/bus-handoff approaches.
+- Links the guidance to a 2026-05-14 hygiene-history tick shard.
+
+## Review threads
+
+### Thread 1: .claude/rules/claim-acquire-before-worktree-work.md:65 (unresolved)
+
+**@copilot-pull-request-reviewer** (2026-05-14T23:57:43Z):
+
+This fallback is not actionable for the exact failure described above. Git refuses to check out the same branch in a second worktree regardless of the new path, so `git worktree add <different-path> <branch>` will hit the same branch-in-use error unless the instruction specifies a detached checkout, a new local branch, or stale-worktree pruning.
+
+### Thread 2: .claude/rules/claim-acquire-before-worktree-work.md:74 (unresolved)
+
+**@copilot-pull-request-reviewer** (2026-05-14T23:57:44Z):
+
+`worktree-handoff` is not a supported bus topic/envelope in the current tooling: the bus `Topic` union and TTL map do not include it, and `bus.ts publish` rejects unknown topics. Agents following this rule cannot post the requested envelope until the schema/tooling is added or this is rewritten to use an existing topic and payload.
+
+### Thread 3: .claude/rules/claim-acquire-before-worktree-work.md:83 (unresolved)
+
+**@copilot-pull-request-reviewer** (2026-05-14T23:57:44Z):
+
+The cited tick shard does not identify this as Otto-Desktop's worktree; it describes a stale `/private/tmp/zeta-otto-id-alloc` entry and attributes the already-pushed fix to another agent/Aaron+Co-Authored-Claude. Reword the empirical anchor to match the cited source, or add a citation that supports the Otto-Desktop ownership claim.
+
+### Thread 4: .claude/rules/claim-acquire-before-worktree-work.md:71 (unresolved)
+
+**@copilot-pull-request-reviewer** (2026-05-14T23:57:44Z):
+
+Use the verb phrase “check out” here; “checkout” is the noun/adjective form for the Git command or target state.