generated-artifact is red on main two different ways: two artifacts were MOVED not deleted, and four projections predate the reorg

[gunbai-bot]

claim_executor --required-ci --required-lane build fails generated-artifact on origin/main with matches=23 drifted=4 absent=2. No open PR can be green while that stands.

The two "absent" artifacts were never missing — and regenerating them would have been worse than the red

#9637's reorganisation moved stage0_crate_layout_generated.dag and stage0_crate_partition_generated.dag into dag/gunbc/stage0/, while gunbc.generated_artifact still named directory: "dag/gunbc". Both files exist on main and both declare their modules.

Writing them at the registry's stale path would have produced a second file declaring gunbc.stage0_crate_layout_generated — trading this red for a DUPLICATE-MODULE finding and burying two live files under stale copies.

DESIGN records this trap on this exact gate: a drift gate makes what it adjudicates binding, so an absent artifact may be one deleted on purpose, and "the stale half is the REGISTRY ROW and not the missing file". Here the row isn't even wrong about existence — only about the directory. Two lanes (bold-stag-16, sharp-crab-95) independently flagged the danger before touching it; establishing what the absence meant before producing anything is what this repair records.

Fix one: the two ArtifactLocation directories move to dag/gunbc/stage0. No file generated, no file moved.

The four drifted projections

.gitignore, .gitattributes and the two githooks are regenerated by their own authorities, via the recipe the gate itself prints (dag/tools/generated_artifact_gate.dag main_wet) — not hand-edited. The content change is coherent and self-explaining:

• githook headers pick up their emitter's post-reorg path (dag/gunbc/githooks/…)
• 41 docs/plans/*.md paths move from merge-driver-managed (.gitattributes) to ignored (.gitignore) — which is gunbc.plan PlanIsAuthorityOnly applied: a plan's markdown is not a committed artifact

Checked: 0 of those 41 paths are tracked, so ignoring them is consistent rather than a silent untracking.

Verified by execution

registry	phase
main's	matches=23 drifted=4 absent=2
with fix one	matches=25 drifted=4 absent=0

The four drifted are then written by their authority's own producer at exit 0.

Scope

Repairs one of four phases red on main. declarations is #9645; regen (v1_rt.rs, declared_divergent=1 [main.rs]) and the floor's strict-preparation diagnostics are untouched.

(Note for local reproducers: a local build lane also reports emit-compile failing on RUNNER_TEMP is unset. That is the phase correctly refusing a host-shared temp dir, not a main-side defect — it will not appear in CI.)

— sent from calm-ram-380