[jonahgabriel]

Summary

Ticket OMN-19081 removes the FALLBACK_IMMUTABLE_GRAIN_PROJECTIONS constant and its associated fallback exemption logic from omnibase_infra. The deletion targets the four entries (HandlerProjectionSessionReplay, HandlerProjectionWorkEvents, node_projection_session_replay, node_projection_work_events) in src/omnibase_infra/runtime/health/projection_apply_flow.py, the fallback_immutable_projections parameter of evaluate_projection_apply_flow, the fallback_exempted_projections field of ModelProjectionApplyFlowVerdict, the fallback note in the dimension detail, and the monitor's pass-through in service_runtime_health_monitor.py. The change spans five files with 60 insertions and 164 deletions. The contract key_grain declaration (shipped in #3943, b1981c9) is now the only source of an exemption.

Context and Trigger
The fallback existed because pre-PR proof on dogfood-101 measured omnimarket 0.4.178 (before key_grain, introduced in 0.4.181) turning both content-addressed exposures DEGRADED. The deletion trigger required a reading of the deployed dev-lane omnimarket at or above 0.4.181 from inside the runtime container. The reading on record is from 2026-09-21 on the .201 dev lane, showing omnimarket 0.4.185 with 45 of 409 contracts declaring key_grain. The omnimarket workspace is 0.4.303. No fresh container reading was taken by this lane (it ran on h202 without access to the .201 container); the 2026-09-21 reading is the one relied on.

Acceptance Mapping

• Undeclared grain is not exempted; it is graded and named under apply_flow_key_grain_unresolved (test test_an_undeclared_grain_is_graded_and_named_not_exempted, and the seam test through the real wiring factory).
• A declared mutable grain is graded (test test_a_declared_mutable_grain_is_graded).
• The two live exemptions HandlerProjectionSessionReplay and HandlerProjectionWorkEvents read HEALTHY on a rising gauge when declared immutable (test test_the_two_live_exemptions_pass_on_their_declaration, parametrised over both).

Local Checks

• 26 tests pass in tests/unit/runtime/health/test_key_grain_from_contract_omn19081.py and tests/integration/test_key_grain_contract_seam_omn19081.py.
• 554 tests pass in a wider health/services selection.
• ruff check and format clean.
• mypy on 41 source files clean.
• pre-commit hooks passed at commit.

Not covered

• The refusal of a constant literal coordinate on a mutable grain naming exposure and field is omnimarket's validator (omnimarket/src/omnimarket/validators/key_grain_declared.py), not in this repo and not changed here.
• No live-lane post-merge reading of the dimension was taken.
• Risk: A deployed runtime still carrying an omnimarket without key_grain would grade the two content-addressed exposures DEGRADED, and a lane probing with --degraded-policy fail would turn red.

Evidence-Ticket: OMN-19081
Evidence-Source: OCC#13244

[jonahgabriel]

Lab: host=h202 lane=approved-omn19081-10072330 head=f01890d00eecd7911aebdfe3f7bb4a6d21aeac09 command="uv run pytest tests/unit/runtime/health/test_key_grain_from_contract_omn19081.py tests/integration/test_key_grain_contract_seam_omn19081.py -n 0" observed="26 passed in 2.03s"

[github-actions]

Hostile Reviewer — adversarial findings (OMN-17492)

Models succeeded: qwen3-review, gpt-oss-review
Models failed: none
New finding threads: 0
Deduped (already posted on this PR): 0
Below quorum (one model only, reported not threaded): 10
Nit-level findings suppressed: 0

The model is the FINDER, never the gate: merge is gated only by the
deterministic Hostile Review Thread Gate, which blocks while
hostile-reviewer threads are unresolved. Resolve each thread after
addressing (or rejecting, with a reply) its finding.

Below quorum: 10 finding(s) raised by one model only (OMN-18479)

These are reported and NOT dropped, but they get no thread and do not block: a single model's finding no other model reproduced is not evidence enough to stop a merge. Read them; act on them if they are right.

• [CRITICAL] src/omnibase_infra/runtime/health/projection_apply_flow.py (qwen3-review) — Removal of fallback exemption causes false-positive DEGRADED on pre-key_grain runtimes | The diff removes the FALLBACK_IMMUTABLE_GRAIN_PROJECTIONS constant and the fallback_immutable_projections param
• [MAJOR] src/omnibase_infra/services/service_runtime_health_monitor.py (qwen3-review) — No migration path or version gate for the removed fallback | The diff removes the fallback mechanism entirely and references OMN-19093 as the tracking ticket for deletion, but provides no runtime vers
• [MAJOR] tests/unit/runtime/health/test_key_grain_from_contract_omn19081.py (qwen3-review) — Test suite no longer covers the pre-key_grain runtime scenario | The removed test test_the_0_4_178_replay_reads_healthy specifically validated that a runtime predating key_grain (where grain resolves
• [MAJOR] src/omnibase_infra/runtime/health/projection_apply_flow.py (gpt-oss-review) — Signature change of evaluate_projection_apply_flow may break callers | The function evaluate_projection_apply_flow no longer accepts the fallback_immutable_projections parameter. Existing callers outs
• [MAJOR] src/omnibase_infra/runtime/health/projection_apply_flow.py (gpt-oss-review) — Health check regression for pre‑declaration runtimes | Removing the interim fallback list eliminates exemption for exposures whose contract does not resolve a key grain on runtimes predating the key_g
• [MINOR] src/omnibase_infra/models/health/model_projection_apply_flow_verdict.py (qwen3-review) — Model field removal is a breaking change for any consumer of the verdict | The field fallback_exempted_projections is removed from ModelProjectionApplyFlowVerdict. Any external consumer that deseriali
• [MINOR] src/omnibase_infra/runtime/health/projection_apply_flow.py (qwen3-review) — Docstring in projection_apply_flow.py is now internally inconsistent | The module-level docstring in projection_apply_flow.py was updated to say 'The exposure's contract declares that grain (key_grain
• [MINOR] tests/unit/runtime/health/test_key_grain_from_contract_omn19081.py (qwen3-review) — No test verifies that the two live projections are actually declared immutable in their contracts | The new test test_the_two_live_exemptions_pass_on_their_declaration parametrizes over HandlerProject
• [MINOR] tests/integration/test_key_grain_contract_seam_omn19081.py (gpt-oss-review) — Missing test for older runtime version scenario | The test suite no longer includes a case that verifies health behavior when the runtime's omnimarket predates the key_grain declaration, leaving the n
• [MINOR] src/omnibase_infra/runtime/health/projection_apply_flow.py (gpt-oss-review) — Outdated docstring references to removed fallback parameter | The docstring for evaluate_projection_apply_flow still describes the fallback_immutable_projections argument, which no longer exists, caus

[github-actions]

✅ Hostile Reviewer — REVIEWED

Critical findings: 1
Major findings: 4
Total findings: 10
Models succeeded: qwen3-review,gpt-oss-review
Models unavailable: none

Action required: findings were posted as review threads. Address or reject each one, then resolve its thread — the Hostile Review Thread Gate fails while hostile-reviewer threads are unresolved (OMN-17492).

---

Semantics (OMN-17492 — the model finds, thread resolution gates)

Surface	Meaning	Blocks merge?
Review threads	Per-finding, posted by the reviewer	No (informational)
Hostile Review Thread Gate	Deterministic: unresolved hostile-reviewer threads exist	Fails until resolved (not yet a required context)
degraded verdict	Fewer than 2 models succeeded (infra)	Fails this job with a named reason

Powered by omniintelligence.review_pairing.cli_review — multi-model adversarial review: qwen3-review (Qwen3.8-27B), gpt-oss-review (gpt-oss-120b) (OMN-8468/OMN-8524/OMN-17492)