fix: treat patch-equivalence to default as a merged cleanup signal

[chubes4]

Root cause

workspace worktree cleanup / active-no-signal-merged-apply decided a branch was "merged to default" primarily via git ancestry (rev-list --count origin/main..branch). Squash merges create a brand-new commit on origin/main, so the original branch tip is never an ancestor of origin/main. Every squash-merged worktree was therefore classified not_merged_to_default and never auto-cleaned.

The tooling already computed patch-equivalence (git cherry) and emitted the label patch_equivalent_default, but then treated it as a skip reason instead of a removal signal.

What changed

WorkspaceWorktreeCleanupEngine::detect_local_merged_signal()

After the ancestry check fails, the method now falls back to git cherry <default_ref> <branch_ref>. If every local commit has a patch-equivalent match on the remote default branch, it returns a new signal patch-equivalent-merged and marks preserve_local_branch: true (the commits are equivalent but not contained, so we keep the local branch rather than force-delete it).

WorktreeCleanupCandidateClassifier::classify_merge_signal_path()

Passes preserve_local_branch from the signal into the candidate row so the removal loop skips local-branch deletion for squash-merged worktrees.

WorkspaceActiveNoSignalCleanup

• suggest_active_no_signal_action() now evaluates dirty/unpushed and open-PR guards before patch-equivalence, so an open PR or dirty worktree cannot be classified patch_equivalent_default.
• worktree_active_no_signal_merged_apply() accepts patch_equivalent_default rows and dispatches to a new metadata builder that revalidates current state via git cherry.
• New build_current_patch_equivalent_to_default_cleanup_evidence() validates clean/unpushed and re-proves patch-equivalence before writing STATE_MERGED cleanup metadata.

Safety guards kept intact

• Dirty worktrees → still unsafe_dirty_or_unpushed / dirty_worktree.
• Unpushed commits → still unpushed_commits.
• Open PRs → still active_open_pr.
• The new patch-equivalent-merged signal only fires after the existing dirty/unpushed probes have passed.

GitHub PR state

Not wired into the default cleanup path. The fix is intentionally pure-git (git cherry) and does not require a GitHub API call, matching the issue's request to prefer the self-contained fix as primary.

Verification

• homeboy review data-machine-code lint --changed-since main → passed (0 PHPCS/PHPStan findings).
• homeboy review data-machine-code test --changed-since main → test runner could not start because the local WP Codebox CLI is not configured (wp-codebox --version failed). This is an environment bootstrap issue, not a code issue.
• Native repo tests: php tests/*.php all pass, including the extended tests/worktree-cleanup-patch-equivalence.php which now asserts:
  • squash-equivalent branch produces patch-equivalent-merged signal,
  • clean/no-PR patch-equivalent rows suggest patch_equivalent_default,
  • open PR blocks to active_open_pr,
  • dirty worktree blocks to unsafe_dirty_or_unpushed,
  • active_no_signal_merged_apply plans patch_equivalent_default rows.

Closes #882

[homeboy-ci]

Homeboy Results — data-machine-code

Lint

✅ review lint — passed

ℹ️ Full options: homeboy self docs commands/lint
Deep dive: homeboy review lint data-machine-code --changed-since b6a9525

Artifacts and drill-down

• CI results artifact: homeboy-ci-results-data-machine-code-review-lint-quality-Linux-node24 contains immediate command JSON for this action invocation.
• Observation artifact: homeboy-observations-data-machine-code-review-lint-quality-Linux-node24 contains exported Homeboy run history for deeper queries.
• Drill-down: download the observation artifact, then run homeboy runs import <dir>, homeboy runs list, and homeboy runs findings <run-id>.
• Artifacts are attached to the workflow run: https://github.com/Extra-Chill/data-machine-code/actions/runs/28955873646

Test

✅ review test — passed

ℹ️ Auto-fix lint issues: homeboy refactor data-machine-code --from lint --write
ℹ️ Collect coverage: homeboy test data-machine-code --coverage
ℹ️ Save test baseline: homeboy test data-machine-code --baseline
ℹ️ Pass args to test runner: homeboy test -- [args]
ℹ️ Full options: homeboy self docs commands/test
Deep dive: homeboy review test data-machine-code --changed-since b6a9525

Artifacts and drill-down

• CI results artifact: homeboy-ci-results-data-machine-code-review-test-quality-Linux-node24 contains immediate command JSON for this action invocation.
• Observation artifact: homeboy-observations-data-machine-code-review-test-quality-Linux-node24 contains exported Homeboy run history for deeper queries.
• Drill-down: download the observation artifact, then run homeboy runs import <dir>, homeboy runs list, and homeboy runs findings <run-id>.
• Artifacts are attached to the workflow run: https://github.com/Extra-Chill/data-machine-code/actions/runs/28955873646

Audit

✅ review audit — passed

• audit — 9 finding(s)
• Total: 9 finding(s)

Deep dive: homeboy review audit data-machine-code --changed-since b6a9525

Artifacts and drill-down

• CI results artifact: homeboy-ci-results-data-machine-code-review-audit-quality-Linux-node24 contains immediate command JSON for this action invocation.
• Observation artifact: homeboy-observations-data-machine-code-review-audit-quality-Linux-node24 contains exported Homeboy run history for deeper queries.
• Drill-down: download the observation artifact, then run homeboy runs import <dir>, homeboy runs list, and homeboy runs findings <run-id>.
• Artifacts are attached to the workflow run: https://github.com/Extra-Chill/data-machine-code/actions/runs/28955873646

Tooling versions

• Homeboy CLI: homeboy 0.281.20+123145ee082b+9702c22
• Extension: wordpress from https://github.com/Extra-Chill/homeboy-extensions
• Extension revision: f89b4297
• Action: unknown@unknown