Gate AIGuard max queue raw context

[hyeokjun32]

Summary

• require AIGuard runtime_queue_overload raw context to preserve orchestrator_candidate_operation_max_total_queue_depth=7
• update the Runtime Intelligence chain guard fixture and smoke assertions
• document the Lab bundle manifest raw-context traceability gate

Tests

• python -m pytest -q tests/test_runtime_intelligence_bundle_manifest.py tests/test_runtime_intelligence_evidence_chain_smoke.py -p no:cacheprovider
• python -m pytest -q -p no:cacheprovider
• python -m scripts.check_runtime_intelligence_bundle_manifest --manifest examples/runtime_intelligence_chain/bundle_manifest.json --summary-out /private/tmp/inferedgelab_bundle_manifest_max_queue_raw_context.md
• poetry run inferedgelab portfolio-demo-check
• bash scripts/smoke_runtime_intelligence_chain.sh --output-dir /private/tmp/inferedgelab_runtime_intelligence_max_queue_raw_context_20260601
• git diff --check