[jonahgabriel]

Ticket: OMN-15645

This change addresses the catalog worker omission by hardcoding DELEGATION_ROUTING_TIERS_PATH=/app/config/delegation/routing_tiers.yaml in docker/catalog/services/runtime-worker.yaml. This explicit deployment pin matches other runtime kernels and resolves the RED cause: "runtime/runtime-worker must bind DELEGATION_ROUTING_TIERS_PATH to /app/config/delegation/routing_tiers.yaml; got None".

The falsified premise that an unset key causes a current boot failure is corrected: on live h201 worker omnimarket 0.4.307, a separate process with the key blank called real _get_config() and exited 0, using packaged bootstrap tiers for blank/unset keys. The remaining omission was that CatalogResolver and generate_compose rendered runtime-worker with DELEGATION_ROUTING_TIERS_PATH absent for both runtime and runtime-infrastructure bundles with RUNTIME_PROFILE=workers.

Lab: host=h201 lane=lab-fill-omn15645-10090550 command=uv run --frozen pytest tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py -q --no-cov observed=15 passed

Read-only live loader proof: docker exec using the inherited deployment environment on omnibase-infra-runtime-worker-1 loaded ModelDelegationConfig; bound file exists and SHA256 matches packaged omnimarket copy: 5a54e42d851e911275a58e7f7177990db2f03aa838cabf0533a6b5b7886689da.

The focused tests changed from 2 failed and 13 passed to 15 passed. The existing guard now covers workers in both bundles and runs as delegation-routing-tiers-binding before commit and in the compose-required-env-coverage CI job. The new hook, yamlfmt, end-of-file-fixer, Ruff lint and format checks passed.

Config-only Docker Compose renders for dev, stability-test, judge and prod returned exit 0 with the expected pin on each runtime kernel in that profile. Renders used isolated labfill-omn15645-render-* project names, the repository's lane overlays and synthetic test fixture values; judge used its judge profile. They did not recreate a runtime.

Acceptance evidence remains partial: this branch has not had a fresh runtime bring-up, the historical unset-key refusal cannot be reproduced on the current worker, and this lane has not checked the prod image's file in a container. The loader probe used a fresh process in the existing worker with its inherited deployment binding; it did not deploy this branch. Full repository hooks and new-PR CI remain pending. The launching runner must commit, push, open the PR and hand it off.

OMN-15645: Launching runner commit was refused by ci-live-contact because the changed CI workflow lacked recorded contact evidence. Real h201 worker omnibase-infra-runtime-worker-1 was probed using its inherited deployment environment, without an env override. The real _get_config() loaded ModelDelegationConfig; bound and packaged file hashes matched. Captured result and source/capture provenance are stored in tests/fixtures/omn15645_worker_tiers_live_contact.json. The existing catalog binding test now consumes this recording through recorded_response and a pytest.mark.live_contact marker, comparing generated runtime pins with the recorded deployed worker pin. The deployment-binding hook also triggers on changes to the recording. ci-live-contact and delegation-routing-tiers-binding hooks passed; the focused test file has 15 passed; Ruff lint and format checks passed. Five changed files are staged. Commit message is unchanged. The launching runner still owns commit, push, PR creation and handoff. Fresh branch runtime bring-up remains pending.

Evidence-Ticket: OMN-15645
Evidence-Source: OCC#13388

[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): 8
Nit-level findings suppressed: 1

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: 8 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] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (gpt-oss-review) — Fixture assertion contradicts parametrize scope | The test test_catalog_runtime_services_bind_the_packaged_tiers_path is parametrized to cover multiple service names, including omninode-runtime an
• [MAJOR] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (qwen3-review) — Fixture assertion contradicts parametric service coverage | The test test_catalog_runtime_services_bind_the_packaged_tiers_path is parametrized over seven (bundle, service_name) pairs, including `om
• [MAJOR] tests/fixtures/omn15645_worker_tiers_live_contact.json (qwen3-review) — SHA256 hash in fixture is not verified against source of truth | The fixture omn15645_worker_tiers_live_contact.json contains a hardcoded file_sha256 and packaged_sha256 value. The test asserts
• [MAJOR] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (gpt-oss-review) — Hardcoded SHA256 in fixture creates brittle integrity check | The test asserts recorded_response["file_sha256"] == recorded_response["packaged_sha256"]. Both values are hardcoded in the JSON fixture
• [MINOR] .pre-commit-config.yaml (qwen3-review) — Pre-commit hook regex may not cover all relevant files | The pre-commit hook delegation-routing-tiers-binding uses a regex to match files that should trigger the test. The regex includes `docker/(Do
• [MINOR] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (qwen3-review) — Test does not verify that the path is actually baked into the image | The test test_catalog_runtime_services_bind_the_packaged_tiers_path checks that the compose file generated by the catalog resolv
• [MINOR] tests/fixtures/omn15645_worker_tiers_live_contact.json (qwen3-review) — Fixture provenance is not machine-verifiable | The fixture includes a _provenance field with a command that was used to capture the data. This command is a complex one-liner that runs `docker exec
• [MINOR] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (gpt-oss-review) — Missing coverage for 'runtime-infrastructure' bundle in other tests | The new parametrization includes ("runtime-infrastructure", "runtime-worker"). However, the other tests in the file (e.g., `test

[github-actions]

✅ Hostile Reviewer — REVIEWED

Critical findings: 0
Major findings: 3
Total findings: 9
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)

[github-actions]

Hostile Reviewer — adversarial findings (OMN-17492)

Models succeeded: qwen3-review, gpt-oss-review
Models failed: none
New finding threads: 3
Deduped (already posted on this PR): 0
Below quorum (one model only, reported not threaded): 7
Nit-level findings suppressed: 2

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: 7 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.

• [MAJOR] .pre-commit-config.yaml (qwen3-review) — Pre-commit hook regex does not cover the new fixture file | The pre-commit hook delegation-routing-tiers-binding has a files: regex that includes `tests/fixtures/omn15645_worker_tiers_live_contact
• [MAJOR] tests/fixtures/omn15645_worker_tiers_live_contact.json (gpt-oss-review) — Hardcoded SHA256 hash in fixture is brittle and unverifiable | The fixture omn15645_worker_tiers_live_contact.json contains a hardcoded file_sha256 and packaged_sha256 value. The test asserts th
• [MINOR] tests/fixtures/omn15645_worker_tiers_live_contact.json (qwen3-review) — Fixture contains a full shell command with embedded Python | The _provenance.command field in the fixture contains a full docker exec command with an inline Python script. While this is for proven
• [MINOR] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (qwen3-review) — Fixture SHA256 values are not validated against a known-good source | The fixture contains file_sha256 and packaged_sha256 fields that are asserted to be equal in the test. However, there is no in
• [MINOR] contracts/OMN-15645.yaml (qwen3-review) — Contract file claims a scope that the tests do not fully cover | The contract OMN-15645.yaml states that the evidence requirement is 'Catalog resolver and generator tests show main, effects and work
• [MINOR] .pre-commit-config.yaml (gpt-oss-review) — Pre-commit hook regex may not cover all relevant catalog files | The pre-commit hook delegation-routing-tiers-binding uses a regex that includes catalog/.*\.ya?ml but does not explicitly cover `do
• [MINOR] tests/fixtures/omn15645_worker_tiers_live_contact.json (gpt-oss-review) — Test fixture is not validated against the actual deployment environment | The fixture omn15645_worker_tiers_live_contact.json is described as a 'live contact' captured from a real container, but the

Findings demoted from threads (anchor rejected)

• [MAJOR] hostile-reviewer (qwen3-review)

  Test asserts against fixture data instead of the constant | The test test_catalog_runtime_services_bind_the_packaged_tiers_path now asserts that the generated environment variable equals recorded_response["tiers_path"] instead of the module-level constant _EXPECTED_PATH. While the fixture currently contains the same value, this creates a circular dependency: the test validates the catalog output against a recorded fixture rather than the known-correct constant. If the fixture is corrupted or updated w

  Resolve this thread when addressed — the Hostile Review Thread Gate blocks while hostile-reviewer threads are unresolved (OMN-17492).

• [MAJOR] hostile-reviewer (qwen3-review)

  Fixture fixture is required for all parametrized cases, not just worker | The recorded_response fixture is marked with @pytest.mark.live_contact and is injected into every parametrized invocation of test_catalog_runtime_services_bind_the_packaged_tiers_path, including cases for omninode-runtime and runtime-effects in the runtime-core, runtime, and local bundles. The fixture file is specifically named omn15645_worker_tiers_live_contact.json and its provenance states it was captured from a w

  Resolve this thread when addressed — the Hostile Review Thread Gate blocks while hostile-reviewer threads are unresolved (OMN-17492).

• [MAJOR] hostile-reviewer (gpt-oss-review)

  Fixture path assertion contradicts test parametrization | The test test_catalog_runtime_services_bind_the_packaged_tiers_path asserts recorded_response["runtime_profile"] == "workers" for all parametrized cases. However, the parametrization includes omninode-runtime and runtime-effects services, which are not worker profiles. The fixture omn15645_worker_tiers_live_contact.json contains "runtime_profile": "workers". This assertion will fail for non-worker services, causing the test to be incorrec

  Resolve this thread when addressed — the Hostile Review Thread Gate blocks while hostile-reviewer threads are unresolved (OMN-17492).

[jonahgabriel]

Lab: head=272243122 host=h202 lane=cause-omnibase_infra-ci-summary-9f8b
Commit 2722431 adds contracts/OMN-15645.yaml, which repo-evidence / dod-verify required at this head. On h202, tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py gave 15 passed, and 2 failed with the dev docker/catalog/services/runtime-worker.yaml. node_dod_verify on the contract returned status verified.

[github-actions]

Hostile Reviewer — adversarial findings (OMN-17492)

Models succeeded: qwen3-review, gpt-oss-review
Models failed: none
New finding threads: 4
Deduped (already posted on this PR): 0
Below quorum (one model only, reported not threaded): 4
Nit-level findings suppressed: 1

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: 4 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.

• [MAJOR] tests/fixtures/omn15645_worker_tiers_live_contact.json (qwen3-review) — SHA256 hash in fixture is not verified against actual file | The fixture omn15645_worker_tiers_live_contact.json contains a hardcoded SHA256 hash `5a54e42d851e911275a58e7f7177990db2f03aa838cabf0533a
• [MINOR] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (qwen3-review) — Fixture design couples test to specific service type | The recorded_response fixture is designed to read a worker-specific contact file, but is used in a test that covers multiple service types. Thi
• [MINOR] tests/unit/docker/test_delegation_routing_tiers_path_binding_omn15645.py (qwen3-review) — Fixture marker assertion is fragile | The fixture asserts marker is not None and len(marker.args) == 1, but does not validate that the marker argument is a valid file path or that the file exists. I
• [MINOR] tests/fixtures/omn15645_worker_tiers_live_contact.json (gpt-oss-review) — Hardcoded SHA256 in fixture without verification mechanism | The fixture omn15645_worker_tiers_live_contact.json contains hardcoded file_sha256 and packaged_sha256 values. The test asserts these

Findings demoted from threads (anchor rejected)

• [MAJOR] hostile-reviewer (qwen3-review)

  Fixture assertion contradicts test parametrization | The test test_catalog_runtime_services_bind_the_packaged_tiers_path is parametrized across multiple services including omninode-runtime and runtime-effects, but the recorded_response fixture asserts recorded_response["runtime_profile"] == "workers" for every parametrized case. The fixture file omn15645_worker_tiers_live_contact.json contains "runtime_profile": "workers". This assertion will fail for the omninode-runtime and `runtime-effect

  Resolve this thread when addressed — the Hostile Review Thread Gate blocks while hostile-reviewer threads are unresolved (OMN-17492).

• [MAJOR] hostile-reviewer (gpt-oss-review)

  Fixture assertion mismatch for non-worker services | The test test_catalog_runtime_services_bind_the_packaged_tiers_path is parameterized to include omninode-runtime and runtime-effects services, but it asserts that recorded_response["runtime_profile"] == "workers". The fixture omn15645_worker_tiers_live_contact.json contains "runtime_profile": "workers". This assertion will pass for the worker service, but the test logic implies that the fixture is being used to validate the environment binding

  Resolve this thread when addressed — the Hostile Review Thread Gate blocks while hostile-reviewer threads are unresolved (OMN-17492).

• [MINOR] hostile-reviewer (qwen3-review)

  Missing test coverage for local bundle worker service | The test parametrization includes ("runtime", "runtime-worker") and ("runtime-infrastructure", "runtime-worker"), but does not include ("local", "runtime-worker"). The local bundle is included for omninode-runtime and runtime-effects, but not for runtime-worker. This is an incomplete test matrix. If the local bundle also includes a runtime-worker service, it should be tested. If it does not, the omission should be documented. | Eviden

  Resolve this thread when addressed — the Hostile Review Thread Gate blocks while hostile-reviewer threads are unresolved (OMN-17492).

• [MINOR] hostile-reviewer (gpt-oss-review)

  Missing test for 'local' bundle worker service | The parameterized test includes ("runtime", "runtime-worker") and ("runtime-infrastructure", "runtime-worker") but does not include ("local", "runtime-worker"). The local bundle is tested for omninode-runtime and runtime-effects, but not for runtime-worker. If the local bundle is intended to support the worker service, this is a missing test case. If it is not, the omission is not documented. | Evidence: @pytest.mark.parametrize(
  ("bundle"

  Resolve this thread when addressed — the Hostile Review Thread Gate blocks while hostile-reviewer threads are unresolved (OMN-17492).