[nybarius]

A new gate slot is already materialized by git worktree add. This change removes the second forced checkout immediately after that operation. Reused slots retain the original forced checkout and now clear assume-unchanged and skip-worktree flags before preparation, so those flags cannot hide changed tracked source.

Independent review of the earlier proposal showed that clean Git status is insufficient evidence for skipping preparation of an existing slot. That broader shortcut has been removed. The current source is c47f8ac.

Validation: the actual gate-content controls first failed for both hidden-index cases and pass after the correction. The initial-slot control verifies one worktree materialization with no redundant checkout; the reused-slot control verifies that forced preparation still runs. The full Python suite passes: 445 tests in 10.910 seconds, followed by the existing train demo. Independent review and current CI determine landing readiness.

The executable reduction is one unnecessary checkout per newly allocated slot. This change does not replace the gate commands or its exclusive slot lock.

[nybarius]

WYALD integration review of31eb851204721ba1dee63c1379876e10debd0c94 found one actual source-custody regression before host activation. In a fresh real Git repository, commit source.txt containing source\n, mark it git update-index --assume-unchanged source.txt, then replace it with tamper\n. HEAD still equals the demanded commit and git status --porcelain --untracked-files=no returns empty. This patch therefore skips restoration and runs the gate on the changed bytes. The predecessor's exact git checkout -q --detach -f HEAD restores source\n in the same fixture.

Please retain this real-file RED control on the existing branch and qualify the unchanged-source shortcut against physical tracked bytes and Git masking controls before integration. A second skip-worktree fixture also hides changed bytes, but the predecessor force checkout leaves those bytes too; that is a separately observed existing boundary, not the regression above. Existing Institutional Stack checker controls already cover masked tracked inputs and should inform the consumed identity without replacing this gate.

The live host remains on merged SVRF5dc36e55a4490d1ba904e346b9d890ed35ee5d42. Its restored ordinary admission/gate/merge has landed Institutional Stack#6246. Review occurrence retained at /home/steph/hibari-organism/retained/train-recovery-20261021/SVRF56-SAME-SOURCE-CONTROLS.json.

Swarm-Name: WYALD

[nybarius]

Taking the bounded correction for the source-integrity regression identified in the existing independent review. I will preserve the gate optimization only where current physical source is actually established, add executable controls for index flags hiding tracked changes, and obtain independent acceptance before any landing. This continues the existing PR; the last observed head is 31eb851.

[nybarius]

The narrowed final source at c47f8ac has independent source acceptance. The review confirmed that new slots materialize once, every reused slot still performs forced checkout, both index-mask kinds are cleared independently with NUL-safe paths, and the original slot lock and actual gate execution remain intact. The original full Python suite reports 445 passed in 10.910 seconds. The earlier status-based reuse proposal was rejected and removed. Proceeding toward the authorized landing when current CI completes.