feat(OMN-10580): add CI grep gate for leaked literals with allowlist governance

.github/workflows/auto-merge.yml

@@ -2,7 +2,7 @@
 # SPDX-License-Identifier: MIT
 #
 # OMN-6571: Automatically enable auto-merge (squash) on every PR opened
-# by jonahgabriel. Ensures PRs do not sit idle in the merge queue waiting
+# by jonahgabriel. Ensures PRs do not sit idle in the merge queue waiting  # onex-allow-raw-env OMN-10580 reason="CI workflow scoped to repo owner actor; not a personal credential"
 # for manual enrollment.
 #
 # OMN-7786: Re-trigger on review and check_suite events so the workflow
@@ -77,7 +77,7 @@ jobs:
       - uses: actions/checkout@v4
 
       - name: Enable auto-merge
-        if: steps.resolve.outputs.skip != 'true' && steps.resolve.outputs.actor == 'jonahgabriel'
+        if: steps.resolve.outputs.skip != 'true' && steps.resolve.outputs.actor == 'jonahgabriel'  # onex-allow-raw-env OMN-10580 reason="CI actor gate for repo owner; not a hardcoded credential"
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           PR: ${{ steps.resolve.outputs.pr }}

.github/workflows/linear-close-on-merge.yml

@@ -10,7 +10,7 @@
 # and calls the Linear API to mark the ticket as Done.
 #
 # Branch name conventions supported:
-#   jonahgabriel/omn-XXXX-description  (Linear default)
+#   jonahgabriel/omn-XXXX-description  (Linear default)  # onex-allow-raw-env OMN-10580 reason="branch name format example in comment; not a credential or connection string"
 #   jonah/omn-XXXX-description          (short username variant)
 #   epic/OMN-XXXX/OMN-YYYY/run-id       (epic-team branches, closes child OMN-YYYY)
 #   OMN-XXXX-description                (bare prefix variant)

.github/workflows/pr-arch-review.yml

@@ -23,7 +23,7 @@
 # Bot identity (onexbot GitHub App):
 #   When ONEXBOT_APP_ID + ONEXBOT_APP_PRIVATE_KEY org secrets are set, comments are
 #   posted as onexbot[bot] via a short-lived installation token. Until the GitHub App
-#   is created, the workflow falls back to GITHUB_TOKEN (posts as jonahgabriel).
+#   is created, the workflow falls back to GITHUB_TOKEN (posts as github-actions[bot]).
 #
 #   To create the app:
 #     https://github.com/organizations/OmniNode-ai/settings/apps/new
@@ -33,6 +33,7 @@
 #
 #   bot_login constants in handler_thread_poster.py and handler_thread_watcher.py
 #   must match the actual app slug: "onexbot[bot]"
+#   Fallback identity is now github-actions[bot] — no personal username in the fallback path.
 
 name: PR Architectural Review
 
@@ -80,8 +81,8 @@ jobs:
             echo "Using onexbot GitHub App token"
           else
             echo "token=${{ secrets.GITHUB_TOKEN }}" >> "$GITHUB_OUTPUT"
-            echo "bot_login=jonahgabriel" >> "$GITHUB_OUTPUT"
-            echo "ONEXBOT_APP_ID/ONEXBOT_APP_PRIVATE_KEY not set — falling back to GITHUB_TOKEN (posts as jonahgabriel)"
+            echo "bot_login=github-actions[bot]" >> "$GITHUB_OUTPUT"
+            echo "ONEXBOT_APP_ID/ONEXBOT_APP_PRIVATE_KEY not set — falling back to GITHUB_TOKEN (posts as github-actions[bot])"
           fi
 
   # ============================================================================

.github/workflows/pr-review-bot.yml

@@ -86,7 +86,7 @@ jobs:
             echo "Using onexbot GitHub App token"
           else
             echo "token=${{ secrets.GITHUB_TOKEN }}" >> "$GITHUB_OUTPUT"
-            echo "bot_login=jonahgabriel" >> "$GITHUB_OUTPUT"
+            echo "bot_login=jonahgabriel" >> "$GITHUB_OUTPUT"  # onex-allow-raw-env OMN-10580 reason="CI fallback bot login set to repo owner when GitHub App not configured"
             echo "ONEXBOT_APP_ID/ONEXBOT_APP_PRIVATE_KEY not set — falling back to GITHUB_TOKEN"
           fi
 

.github/workflows/reject-leaked-literals.yml

@@ -1,21 +1,25 @@
 # SPDX-FileCopyrightText: 2026 OmniNode.ai Inc.
 # SPDX-License-Identifier: MIT
 #
-# OMN-10554: leaked-literals gate, ADVISORY MODE.
+# OMN-10580 (hardens OMN-10554): leaked-literals gate — BLOCKING MODE.
 #
-# Wave 0 of the Public-Shippable plan
-# (`docs/plans/2026-05-05-omnimarket-public-shippable.md`).
+# This workflow runs `scripts/validation/check_leaked_literals.sh blocking all`
+# on every PR and on merge_group entry. Any unannotated leaked literal causes
+# exit 1 and blocks merge.
 #
-# This workflow surfaces the leaked-literals baseline for situational
-# awareness. It runs `scripts/validation/check_leaked_literals.sh advisory all`
-# on every PR and on merge_group entry, prints findings to the run log, and
-# ALWAYS exits 0. It is NOT a required status check.
+# To permit a legitimate occurrence, add an inline annotation on the same line:
+#   # onex-allow-internal-ip OMN-XXXXX reason="<concrete reason>"
+#   # onex-allow-local-path OMN-XXXXX reason="<concrete reason>"
+#   # onex-allow-test-fixture OMN-XXXXX reason="<concrete reason>"
+#   # onex-allow-raw-env OMN-XXXXX reason="<concrete reason>"
+#   # onex-allow-model-id OMN-XXXXX reason="<concrete reason>"
 #
-# Wave 3 (Task 8 blocking-mode) flips MODE to `blocking` after Tasks 3+7 land
-# their scrubs and brings this workflow under "required status checks". Until
-# then, finding count is informational and downstream tasks consume the report.
+# See docs/leaked-literals-governance.md for the full governance policy.
+#
+# This job is a REQUIRED STATUS CHECK. The job name
+# "Leaked Literals Gate" must match branch protection rules.
 
-name: Leaked Literals Gate (advisory)
+name: Leaked Literals Gate
 
 on:
   pull_request:
@@ -27,15 +31,15 @@ permissions:
   contents: read
 
 jobs:
-  leak-gate-advisory:
-    name: Scan for leaked literals (advisory)
+  leak-gate:
+    name: Leaked Literals Gate
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
         uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
-      - name: Run leak gate in advisory mode
+      - name: Run leak gate in blocking mode
         run: |
-          bash scripts/validation/check_leaked_literals.sh advisory all
+          bash scripts/validation/check_leaked_literals.sh blocking all

.leaked-literals-allowlist.yaml

@@ -0,0 +1,214 @@
+# Leaked Literals Allowlist — OMN-10580
+#
+# This file is the machine-readable registry of all intentional
+# leaked-literal exemptions in this repo.
+#
+# Governance: docs/leaked-literals-governance.md
+# Gate script: scripts/validation/check_leaked_literals.sh
+#
+# Format:
+#   - file: <repo-relative path>
+#     type: per_line | file_level | self_exempt
+#     ticket: OMN-XXXXX
+#     reason: <one-line explanation>
+#
+# NOTE: The gate script does NOT parse this file at runtime — exemptions are
+# enforced via inline annotations (# onex-allow-* comments) and the
+# SELF_EXEMPT_FILES array in check_leaked_literals.sh.  This YAML is a
+# human-readable audit trail only.
+
+exemptions:
+  # ── Self-exempt (pattern catalog / CI workflow / this governance doc) ──────
+  - file: scripts/validation/check_leaked_literals.sh
+    type: self_exempt
+    ticket: OMN-10580
+    reason: contains the LEAK_REGEX pattern catalog; self-referential by necessity
+
+  - file: .github/workflows/reject-leaked-literals.yml
+    type: self_exempt
+    ticket: OMN-10580
+    reason: CI workflow that invokes the gate script; references annotation syntax in docs
+
+  - file: docs/leaked-literals-governance.md
+    type: self_exempt
+    ticket: OMN-10580
+    reason: governance doc showing annotation examples; self-referential
+
+  - file: docs/audits/2026-05-05-raw-env-usage.csv
+    type: self_exempt
+    ticket: OMN-10580
+    reason: generated audit report containing literal inventory as data rows
+
+  - file: docs/audits/2026-05-05-raw-env-usage.md
+    type: self_exempt
+    ticket: OMN-10580
+    reason: generated audit summary referencing literal patterns as data
+
+  - file: docs/audits/2026-05-05-contracts-dir-references.csv
+    type: self_exempt
+    ticket: OMN-10580
+    reason: generated audit report; path references are data rows, not source defaults
+
+  - file: docs/tracking/delegation-cost-projection-lane.md
+    type: self_exempt
+    ticket: OMN-10580
+    reason: tracking doc with config examples and lab address as orientation hints
+
+  # ── File-level exemptions (test fixtures) ────────────────────────────────
+  - file: tests/test_emergency_bypass_parser.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses jonahgabriel as synthetic actor in bypass-parser test corpus
+
+  - file: tests/unit/config/test_settings.py
+    type: file_level
+    ticket: OMN-10580
+    reason: pattern catalog in assertions for Settings validation tests
+
+  - file: tests/unit/scripts/test_leak_gate.py
+    type: file_level
+    ticket: OMN-10580
+    reason: writes synthetic leaky files to test the gate script itself
+
+  - file: tests/nodes/node_pr_review_bot/test_workflow_runner_wiring.py
+    type: file_level
+    ticket: OMN-10580
+    reason: LLM wiring test using lab endpoints as allowed-endpoint inputs
+
+  - file: tests/test_webhook_reconciler.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab git actor handle as synthetic PR author in webhook test fixtures
+
+  - file: tests/unit/scripts/test_run_delegation_cost_projection_process.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab Kafka address as monkeypatched env var in delegation cost tests
+
+  - file: tests/unit/nodes/node_overnight/test_main_publish_events.py
+    type: file_level
+    ticket: OMN-10580
+    reason: sets lab Kafka bootstrap address via monkeypatch to test event publishing
+
+  - file: tests/unit/experiments/adk_eval/harness/test_aggregator.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint and model ID as test inputs to verify aggregator wiring
+
+  - file: tests/nodes/node_ab_inference_effect/test_handler.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint as allowed-endpoint test input in request fixtures
+
+  - file: tests/test_session_bootstrap_rev7.py
+    type: file_level
+    ticket: OMN-10580
+    reason: asserts /Users/jonah does NOT appear in prompts; literal appears in assertion string
+
+  - file: tests/test_model_router_routes_to_local_first.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint as router test input to verify local-first routing
+
+  - file: tests/test_golden_chain_local_supervisor.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in routing fixtures for local supervisor golden chain
+
+  - file: tests/nodes/node_merge_sweep_triage_orchestrator/test_phase2_command_models.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab endpoint as command model test input
+
+  - file: tests/integration/test_cost_event_publisher_kafka.py
+    type: file_level
+    ticket: OMN-10580
+    reason: integration test using lab Kafka bootstrap address as test target
+
+  - file: tests/conftest.py
+    type: file_level
+    ticket: OMN-10580
+    reason: lab endpoints used as fixture defaults for golden chain tests
+
+  - file: tests/test_model_router_exponential_backoff.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in router backoff test inputs
+
+  - file: tests/test_model_router_fallback_on_degradation.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in router fallback test inputs
+
+  - file: tests/test_model_router_health_check_degrades.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in router health-check test inputs
+
+  - file: tests/test_model_router_publishes_degradation_event.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in degradation event test inputs
+
+  - file: tests/test_runtime_sha_verify_handler.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab endpoint as SHA verify handler test input
+
+  - file: tests/test_golden_chain_nightly_loop_controller.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in nightly loop controller test fixtures
+
+  - file: tests/test_golden_chain_pr_lifecycle_upgrade.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in PR lifecycle upgrade test fixtures
+
+  - file: tests/test_golden_chain_review_thread_reconciler.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint in review thread reconciler test fixtures
+
+  - file: tests/nodes/node_ci_fix_effect/test_handler_ci_fix.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint as CI fix effect test input
+
+  - file: tests/nodes/node_thread_reply_effect/test_handler_thread_reply.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab git actor handle as synthetic PR branch name in test fixtures
+
+  - file: tests/test_model_router_authorization_and_ttl.py
+    type: file_level
+    ticket: OMN-10580
+    reason: uses lab LLM endpoint as router authorization/TTL test input
+
+  # ── Per-line annotations (source files) ──────────────────────────────────
+  # Listed here for discoverability; authoritative annotation is inline in the source.
+
+  - file: scripts/projection_api_server.py
+    type: per_line
+    ticket: OMN-10580
+    reason: env-var fallback to lab Postgres; override via POSTGRES_HOST / POSTGRES_PORT
+
+  - file: src/omnimarket/nodes/node_baseline_capture/handlers/probes/probe_kafka_topics.py
+    type: per_line
+    ticket: OMN-10580
+    reason: env-var fallback to lab Redpanda admin; override via REDPANDA_ADMIN_HOST
+
+  - file: src/omnimarket/nodes/node_baseline_capture/handlers/probes/probe_system_health.py
+    type: per_line
+    ticket: OMN-10580
+    reason: env-var fallbacks to lab service endpoints; overrideable per env var
+
+  - file: src/omnimarket/nodes/node_platform_readiness/handlers/handler_platform_readiness.py
+    type: per_line
+    ticket: OMN-10580
+    reason: env-var fallback SSH target for lab infra; override via ONEX_INFRA_SSH_TARGET
+
+  - file: src/omnimarket/nodes/node_projection_llm_cost/consumer.py
+    type: per_line
+    ticket: OMN-10580
+    reason: env-var fallback to lab Postgres; override via POSTGRES_HOST

.pre-commit-config.yaml

@@ -43,6 +43,16 @@ repos:
   # Requires omnibase_infra at $OMNIBASE_INFRA_PATH or ../omnibase_infra.
   - repo: local
     hooks:
+      # OMN-10580: Leaked-literals gate — blocks hardcoded LAN IPs, personal
+      # paths, private model IDs from being committed. Add an inline annotation
+      # to exempt a legitimate occurrence (see docs/leaked-literals-governance.md).
+      - id: leaked-literals-gate
+        name: Leaked literals gate (OMN-10580)
+        language: system
+        entry: bash scripts/validation/check_leaked_literals.sh blocking diff
+        pass_filenames: false
+        stages: [pre-commit]
+
       - id: topic-naming-lint
         name: Topic Naming Lint
         language: system

docker-compose.projection.yml

@@ -20,7 +20,7 @@ x-projection-defaults: &projection-defaults
   network_mode: host
   env_file: .env
   environment:
-    KAFKA_BROKERS: ${KAFKA_BROKERS:-192.168.86.201:19092}
+    KAFKA_BROKERS: ${KAFKA_BROKERS:-192.168.86.201:19092}  # onex-allow-internal-ip OMN-10580 reason="compose env-var fallback to lab Kafka; KAFKA_BROKERS must be set for non-lab deployments"
     OMNIDASH_ANALYTICS_DB_URL: ${OMNIDASH_ANALYTICS_DB_URL}
     KAFKA_CONSUMER_GROUP: ${KAFKA_CONSUMER_GROUP:-omnimarket-projections-v1}
     PYTHONUNBUFFERED: "1"

docs/audits/2026-05-05-contract-config-audit.md

@@ -38,4 +38,3 @@ These nodes have handler imports that suggest transport usage not declared in `d
 | `node_navigation_history_reducer` | postgres | handler uses ['postgres'] but contract dependencies[] does not declare it |
 | `node_pr_review_bot` | llm | handler uses ['llm'] but contract dependencies[] does not declare it |
 | `node_projection_llm_cost` | kafka,postgres | handler uses ['postgres'] but contract dependencies[] does not declare it |
-