[jonahgabriel]

Ticket: OMN-19860 (delegation_events names the lane and session that issued each delegation), step 2 of the consumer-first order.

Cause

Step 1 (the response decodes caller_lane and session_id before declaring them, the projection stores caller_lane) is released since 0.4.253 and the dev lane runs 0.4.270. onex delegate already writes metadata.caller_lane and the request session_id. The delegate-skill handler never copied them onto the terminal, so the projection had nothing to store: on the lab dev lane 4 of 4,732 delegation_events rows carry a caller_lane, none of them from the onex delegate path (read-only psql, 2026-10-01T01:5xZ).

Change

• ModelDelegateSkillResponse declares caller_lane (the ledger lane pattern) and session_id (canonical UUID string). Both are omitted from the wire when None, so an anonymous caller's terminal is unchanged. A non-UUID session is dropped, never refused, because it is attribution and must not dead-letter a terminal; the projection model keeps its UUID-typed override.
• HandlerDelegateSkill._dispatch_and_build_terminal, the one stamping point OMN-19514 introduced for ticket_id, now stamps ticket, caller lane and session together, so every terminal (completed, refused, timed out, failed, runtime shutdown) carries them. A malformed lane is logged and dropped.

Verification

• Red first: the new tests/unit/nodes/node_delegate_skill_orchestrator/test_delegate_skill_caller_identity_omn19860.py failed 18 of 23 before the change. Focused: it, the OMN-19514 ticket tests, the caller-lane projection tests and the OMN-15503/15469 seams, 69 passed.
• Wire compatibility (OMN-18868) against the released consumer v0.4.270: WIRE_COMPAT_PASS for both changed wire models.
• Lab: /omni:lab-run --via ssh --lane deleg-disposition --repo omnimarket --ref <this branch> placed on a lab pool host at sha 438681d: uv run pytest tests/unit/nodes/node_delegate_skill_orchestrator tests/unit/delegation plus the two quota seams, 2462 passed, 6 skipped, exit 0.
• ruff and mypy --strict on the three changed sources pass; the commit ran the full pre-commit set.
• Dev-lane readback after deploy: count(caller_lane) over rows written after the deploy by onex delegate --caller-lane <lane> must be non-zero, with the pre-deploy count (4) as the control.

Evidence-Ticket: OMN-19860
Evidence-Source: OCC#12044

[github-actions]

✅ Architectural Review — PASSED

Errors: 0
Warnings: 0

What this checks

Rule	Description
ARCH-TOPIC-001	No hardcoded Kafka topic strings in handler code
ARCH-DI-001	No event_bus=None bypass
ARCH-DI-002	No direct Handler instantiation outside workflow_runner/adapters
ARCH-DI-003	No reinvented DI containers
ARCH-TOPIC-002	contract.yaml topics follow `onex.{cmd

---

No architectural violations found.

Static architectural lint — no model inference (OMN-14176).

[github-actions]

✅ Hostile Reviewer — PASSED

Blocking findings (critical/error): 0
Total findings: 10
Models succeeded: qwen3-review,gpt-oss-review

---

Gate semantics

Verdict	Meaning	Blocks merge?
passed	>=2 models succeeded, no critical/error findings	No
blocked	CRITICAL or ERROR findings found	Yes
degraded	Fewer than 2 models succeeded (infra unavailable/timeout)	Yes (OMN-15110)

Powered by omniintelligence.review_pairing.cli_review — node-based adversarial review via HandlerLlmCliSubprocess (OMN-8524)