IU Core Worker Live — 07 Next macro package

Three candidate next macros. Each is independently approvable; they have a natural order but no hard dependency except where noted.

Option A — First durable structure-op (RECOMMENDED)

Build fn_iu_structure_op_plan/_apply/_verify/_rollback (or a live PostgreSQL-backed StructureStore) so structure_ops_workflow.py drives a real operation. Start with reparent_piece on the flat 158-IU tree — the lowest-risk op (no IU creation, no birth gate). One durable structure-op: plan (dry-run, iu_structure_operation planned) → sovereign approval → apply (enacted, adjacency + iu_relation + iu_tree_path refresh) → verify → keep or rollback.

• Effort: medium-high. Backup + master gate + sovereign approval for apply.
• Unblocks: organic IU events (a structure-op emits STRUCTURE_OP_EVENTS) → the worker records organic iu_route_attempt dry_run rows → clears dry_run=false readiness blocker 3.

Option B — Auto-cut integration proof

Wire the orchestrator iu_core_record phase (gated, post-lifecycle_enact) that drives autocut_bridge.py requests through StructureOpsWorkflow. Needs Option A first (the live StructureStore). Proof: a dry-run cut → bridge maps the result → structure-op plan rows, no apply. No production CUT.

• Effort: medium. Depends on Option A.

Option C — dry_run=false controlled delivery

The full delivery macro (doc 06). Build a DeliveryTarget, re-author the SQL fn_iu_route_deliver dispatcher, relax iu_outbound_route_safe_chk under backup, one controlled dry_run=false delivery. Best after Option A supplies organic dry-run evidence.

• Effort: high. Most blockers (doc 06).

Recommended sequence

A → (B and C in parallel). Option A is the smallest safe step that produces organic events, which both B and C want. Each macro: discover-first, fresh backup, 8 gates, BEGIN…ROLLBACK probe before durable mutation, exact deferral if a gate fails.