{
  "mode": "offline-synthetic-action-identity",
  "baseline_commit": "dd6a63325e149a287e3600d0ae9b7bd2789a83dd",
  "candidate": "working tree at invocation; published with its source commit",
  "baseline_source_sha256": {
    "src/apixly_jev_harness/__init__.py": "9999dcc9843dee69b2c68b7721f8c826c89e2d6f93720e1dbae80a020f304a98",
    "src/apixly_jev_harness/backend.py": "f7709b468d3d7c131fb8be42b882e5a0387db0077d1fc07e420153b0447f6754",
    "src/apixly_jev_harness/callbacks.py": "e14fa70328634bf3eae87c027d9efc33bfde3a3caebe77b388537335c7250298",
    "src/apixly_jev_harness/cli.py": "8d9434bf8a8b539163194a84e3f099761d08ef32201f99cbce46c139f3726d94",
    "src/apixly_jev_harness/context.py": "b9d76abb33f8a1ebe8731fe6d5fe6811445132a6046448fc0e2cc871507d50ea",
    "src/apixly_jev_harness/diagnostics.py": "9e8e98dc4b103c1a6687dbd31bd2dd01739309c7e4021b9ecb0f28b6976a0f71",
    "src/apixly_jev_harness/engine.py": "1db8e15b86aa92f593b5354a1a7e8a2e168bd6e0a4c903fd37a5df336053036c",
    "src/apixly_jev_harness/output.py": "94be8b482e4279e29aca128a35caf54744dbe8db8188f14905488f8ddde93a0b",
    "src/apixly_jev_harness/surfaces.py": "65ce73edfe9f6a3c16de373b3a1b85671c3d2cc661f3a50dd906a15721c7d8f1",
    "src/apixly_jev_harness/task.py": "30b9dc766df860f9897032a1f902e4e1a7afbd6eff8c381d85ea9e1013a0bfe7",
    "src/apixly_jev_harness/task_cli.py": "d1ebbd533c59a76a99812f3692cdee80582b0c2ba832f1fa525f6ff951d93e48",
    "src/apixly_jev_harness/verify.py": "cad171c63144a60d898553eb5a3d7dc0dd2189f2b20006d8edbe9f0f5a78b1e0"
  },
  "candidate_source_sha256": {
    "src/apixly_jev_harness/__init__.py": "85370f537f2dda6126ca9a94e75f4dc113cba6ebbafb8126033a78cbf3220432",
    "src/apixly_jev_harness/backend.py": "f7709b468d3d7c131fb8be42b882e5a0387db0077d1fc07e420153b0447f6754",
    "src/apixly_jev_harness/callbacks.py": "e14fa70328634bf3eae87c027d9efc33bfde3a3caebe77b388537335c7250298",
    "src/apixly_jev_harness/cli.py": "8d9434bf8a8b539163194a84e3f099761d08ef32201f99cbce46c139f3726d94",
    "src/apixly_jev_harness/context.py": "b9d76abb33f8a1ebe8731fe6d5fe6811445132a6046448fc0e2cc871507d50ea",
    "src/apixly_jev_harness/diagnostics.py": "9e8e98dc4b103c1a6687dbd31bd2dd01739309c7e4021b9ecb0f28b6976a0f71",
    "src/apixly_jev_harness/engine.py": "1db8e15b86aa92f593b5354a1a7e8a2e168bd6e0a4c903fd37a5df336053036c",
    "src/apixly_jev_harness/output.py": "94be8b482e4279e29aca128a35caf54744dbe8db8188f14905488f8ddde93a0b",
    "src/apixly_jev_harness/surfaces.py": "65ce73edfe9f6a3c16de373b3a1b85671c3d2cc661f3a50dd906a15721c7d8f1",
    "src/apixly_jev_harness/task.py": "24551f23ba54b5486800f6525086b9b0b8047f78eb0a68964f7918c95219e239",
    "src/apixly_jev_harness/task_cli.py": "a49089248076b4c7a9c2a463d911f34c31366ee10243ca6734352d3f12c168e3",
    "src/apixly_jev_harness/verify.py": "074e9f2a80ca41bab426f10f648558968dc24f30ea7dbbddfbd825e6aba8a9ee"
  },
  "runner_sha256": "e5ceb7da2e98aa445479603700c4632923cd7c0a5039d0ba37ba8c06958d4b06",
  "python_version": "3.14.2",
  "cases": {
    "execute_bool_numeric_alias": {
      "mode": "lowlevel",
      "kind": "bool_alias",
      "group": "guard_fix"
    },
    "execute_nested_bool_numeric": {
      "mode": "lowlevel",
      "kind": "nested_alias",
      "group": "guard_fix"
    },
    "execute_dict_key_bool_numeric": {
      "mode": "lowlevel",
      "kind": "key_alias",
      "group": "guard_fix"
    },
    "joint_bind_irreversible_numeric": {
      "mode": "lowlevel",
      "kind": "irreversible",
      "group": "guard_fix",
      "strategy": "joint",
      "guard_phase": "bind"
    },
    "execute_custom_equality_spoof": {
      "mode": "lowlevel",
      "kind": "custom_value",
      "group": "guard_fix"
    },
    "bind_fake_action_class_spoof": {
      "mode": "lowlevel",
      "kind": "fake_action",
      "group": "guard_fix",
      "guard_phase": "bind"
    },
    "execute_action_subclass_spoof": {
      "mode": "lowlevel",
      "kind": "action_subclass",
      "group": "guard_fix"
    },
    "staged_numeric_alias_compatible": {
      "mode": "lowlevel",
      "kind": "numeric_alias",
      "group": "compatibility"
    },
    "joint_numeric_alias_compatible": {
      "mode": "lowlevel",
      "kind": "numeric_alias",
      "group": "compatibility",
      "strategy": "joint"
    },
    "numeric_dict_key_alias_compatible": {
      "mode": "lowlevel",
      "kind": "numeric_key",
      "group": "compatibility"
    },
    "engine_staged_normal": {
      "mode": "engine",
      "kind": "normal",
      "group": "compatibility"
    },
    "engine_joint_normal": {
      "mode": "engine",
      "kind": "normal",
      "group": "compatibility",
      "strategy": "joint"
    },
    "engine_dependent_fresh_unknown": {
      "mode": "engine",
      "kind": "dependent",
      "group": "compatibility",
      "unknown_usage": true
    },
    "unmodified_int_enum_value": {
      "mode": "engine",
      "kind": "int_enum",
      "group": "compatibility_narrowing"
    },
    "unmodified_string_enum_value": {
      "mode": "engine",
      "kind": "string_enum",
      "group": "compatibility_narrowing"
    },
    "unmodified_dict_subclass_value": {
      "mode": "engine",
      "kind": "dict_subclass",
      "group": "compatibility_narrowing"
    },
    "bind_dict_subclass_metadata": {
      "mode": "engine",
      "kind": "metadata_subclass",
      "group": "compatibility_narrowing"
    },
    "unmodified_string_enum_tool_id": {
      "mode": "engine",
      "kind": "enum_id",
      "group": "compatibility_narrowing"
    }
  },
  "compatibility_parity": true,
  "reproduction": "python benchmarks/action_identity.py --output local-results/action-identity-ab.json",
  "scope": "Eighteen fixed synthetic tasks, three paired repeats with alternating arm order. Seven public low-level Task API fixtures manually replace offered/bound action fields or types; bool/number aliases, nested key/list aliases, numeric irreversible metadata and custom/FakeAction/subclass equality must not pass the typed guard. Standard Engine never performs these action mutations and this is not a model-output attack. Six compatibility fixtures preserve staged/joint numeric 1/1.0 and numeric dictionary-key equivalence, normal Engine progression and two fresh dependent actions with unknown scripted usage. Five per-arm negative fixtures record unmodified JSON-compatible enums/dictionary subclasses or metadata previously accepted and now rejected. Native inference and real HTTPTransport are forbidden; actual model/network/UI calls zero if guards stay zero.",
  "negative_results": "The same-action comparison is recursive and adds local guard work; no speed or fee gain is claimed. Exact builtin-value/action/id requirements narrow compatibility: a baseline's unmodified IntEnum, string enum or dictionary subclass can execute correctly, while treatment bind rejects with failed or an execute guard triggers execution_uncertain before the measured tool body. That conservative executed/intent window remains; these synthetic body-zero rows do not prove global absence of side effects or permit reruns. Negative fixtures use distinct per-arm expected outcomes and are excluded from manual-mutation guard quality. Numeric int/float equivalence intentionally remains. This is synthetic API identity consistency, not AI semantic quality, actual UI acceptance, a Python sandbox or a general capability boundary; trusted deepcopy/serialization and caller code can still run hooks.",
  "cost_conversion": "No tariff or savings assumed. Actual model usage is zero only when both guards remain zero; guard hits record unknown actual usage and fail fixtures. Scripted usage is separate from actual usage. Real perf_counter timing includes Task domain materialization/binding/guards/execution or actual Engine admission, scripted selection, effects and private journal flush/fsync; excludes imports/task-fixture setup and post-call analysis/source hashing. Loop-only frozen monotonic values stabilize context budgets without replacing real timing. Returned bytes omit local archive paths and use a fixed exception projection for direct API rejection, not a fabricated Engine terminal. Scalar input/context sizes do not count provider tokens.",
  "summary": {
    "baseline": {
      "runs": 54,
      "correct": 33,
      "groups": {
        "guard_fix": {
          "runs": 21,
          "correct": 0
        },
        "compatibility": {
          "runs": 18,
          "correct": 18
        },
        "compatibility_narrowing": {
          "runs": 15,
          "correct": 15
        }
      },
      "manual_guard_fix_accepted": 21,
      "effects": 57,
      "scripted_decisions": 27,
      "provider_calls": 6,
      "unknown_scripted_usage_rows": 3,
      "execution_uncertain_rows": 0,
      "execution_uncertain_body_zero_rows": 0,
      "custom_eq_calls": 6,
      "class_property_reads": 3,
      "native_inference_attempts": 0,
      "real_network_attempts": 0,
      "raw_private_marker_exposures": 0,
      "selection_input_bytes": 60207,
      "shared_context_bytes": 7932,
      "median_whole_operation_seconds": 0.0007753334939479828,
      "returned_context_bytes": 17583,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      }
    },
    "candidate": {
      "runs": 54,
      "correct": 54,
      "groups": {
        "guard_fix": {
          "runs": 21,
          "correct": 21
        },
        "compatibility": {
          "runs": 18,
          "correct": 18
        },
        "compatibility_narrowing": {
          "runs": 15,
          "correct": 15
        }
      },
      "manual_guard_fix_accepted": 0,
      "effects": 21,
      "scripted_decisions": 27,
      "provider_calls": 6,
      "unknown_scripted_usage_rows": 3,
      "execution_uncertain_rows": 9,
      "execution_uncertain_body_zero_rows": 9,
      "custom_eq_calls": 0,
      "class_property_reads": 0,
      "native_inference_attempts": 0,
      "real_network_attempts": 0,
      "raw_private_marker_exposures": 0,
      "selection_input_bytes": 60207,
      "shared_context_bytes": 7932,
      "median_whole_operation_seconds": 0.0008155834948411211,
      "returned_context_bytes": 18918,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      }
    }
  },
  "rows": [
    {
      "case": "execute_bool_numeric_alias",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "bool_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_integer"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005208330112509429,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "execute_bool_numeric_alias",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "bool_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0010131669987458736,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "execute_bool_numeric_alias",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "bool_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0009805840090848505,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "execute_bool_numeric_alias",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "bool_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_integer"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007443749927915633,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "execute_bool_numeric_alias",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "bool_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_integer"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007913750014267862,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "execute_bool_numeric_alias",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "bool_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007699580019107088,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "execute_nested_bool_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "nested_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007752079982310534,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "execute_nested_bool_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "nested_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007893750007497147,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "execute_nested_bool_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "nested_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007695419917581603,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "execute_nested_bool_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "nested_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007364590128418058,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "execute_nested_bool_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "nested_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005727920070057735,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "execute_nested_bool_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "nested_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000785625001299195,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "execute_dict_key_bool_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "key_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007251670031109825,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "execute_dict_key_bool_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "key_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007260830025188625,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "execute_dict_key_bool_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "key_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008372089941985905,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "execute_dict_key_bool_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "key_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0008219999872380868,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "execute_dict_key_bool_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "key_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007503749948227778,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "execute_dict_key_bool_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "key_alias",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007560840022051707,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "joint_bind_irreversible_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "irreversible",
        "group": "guard_fix",
        "strategy": "joint",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007305830076802522,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "joint_bind_irreversible_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "irreversible",
        "group": "guard_fix",
        "strategy": "joint",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_offered",
        "error_type": "ValueError",
        "stopped_guard_phase": "bind",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005667499935952947,
      "returned_context_bytes": 80,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "joint_bind_irreversible_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "irreversible",
        "group": "guard_fix",
        "strategy": "joint",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_offered",
        "error_type": "ValueError",
        "stopped_guard_phase": "bind",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005467500013764948,
      "returned_context_bytes": 80,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "joint_bind_irreversible_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "irreversible",
        "group": "guard_fix",
        "strategy": "joint",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005383329989854246,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "joint_bind_irreversible_numeric",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "irreversible",
        "group": "guard_fix",
        "strategy": "joint",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006826250028098002,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "joint_bind_irreversible_numeric",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "irreversible",
        "group": "guard_fix",
        "strategy": "joint",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_offered",
        "error_type": "ValueError",
        "stopped_guard_phase": "bind",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005491249903570861,
      "returned_context_bytes": 80,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "execute_custom_equality_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "custom_value",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "unsupported_custom_value"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 1,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006939579907339066,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "execute_custom_equality_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "custom_value",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005902920092921704,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "execute_custom_equality_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "custom_value",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005644170014420524,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "execute_custom_equality_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "custom_value",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "unsupported_custom_value"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 1,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005541249993257225,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "execute_custom_equality_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "custom_value",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "unsupported_custom_value"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 1,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005361669900594279,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "execute_custom_equality_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "custom_value",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.00055291700118687,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "bind_fake_action_class_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "fake_action",
        "group": "guard_fix",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 1,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005322499928297475,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "bind_fake_action_class_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "fake_action",
        "group": "guard_fix",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_offered",
        "error_type": "ValueError",
        "stopped_guard_phase": "bind",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005361250077839941,
      "returned_context_bytes": 80,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "bind_fake_action_class_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "fake_action",
        "group": "guard_fix",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_offered",
        "error_type": "ValueError",
        "stopped_guard_phase": "bind",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005127089971210808,
      "returned_context_bytes": 80,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "bind_fake_action_class_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "fake_action",
        "group": "guard_fix",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 1,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005270000110613182,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "bind_fake_action_class_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "fake_action",
        "group": "guard_fix",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 1,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005321669887052849,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "bind_fake_action_class_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "fake_action",
        "group": "guard_fix",
        "guard_phase": "bind",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_offered",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_offered",
        "error_type": "ValueError",
        "stopped_guard_phase": "bind",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007189589960034937,
      "returned_context_bytes": 80,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "execute_action_subclass_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "action_subclass",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 1,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007219999970402569,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "execute_action_subclass_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "action_subclass",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005194580007810146,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "execute_action_subclass_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "action_subclass",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005679160094587132,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "execute_action_subclass_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "action_subclass",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 1,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007397920126095414,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "execute_action_subclass_spoof",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "action_subclass",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 1,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005739170010201633,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "execute_action_subclass_spoof",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "action_subclass",
        "group": "guard_fix",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "rejected",
        "effects": 0,
        "steps": null,
        "scripted_decisions": 0,
        "code": "action_not_bound",
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "rejected",
        "complete": null,
        "reason": null,
        "code": "action_not_bound",
        "error_type": "ValueError",
        "stopped_guard_phase": "execute",
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007472500001313165,
      "returned_context_bytes": 78,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "staged_numeric_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005660840106429532,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "staged_numeric_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007939579954836518,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "staged_numeric_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005571249930653721,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "staged_numeric_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005466250004246831,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "staged_numeric_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005837909993715584,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "staged_numeric_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005636249989038333,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "joint_numeric_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005459170060930774,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "joint_numeric_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007577079959446564,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "joint_numeric_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005680419999407604,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "joint_numeric_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007208329916466027,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "joint_numeric_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005398339999374002,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "joint_numeric_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_alias",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_float"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000726458994904533,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "numeric_dict_key_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_key",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005819590005557984,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "numeric_dict_key_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_key",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005864580016350374,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "numeric_dict_key_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_key",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007641669944860041,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "numeric_dict_key_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_key",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005715000006603077,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "numeric_dict_key_alias_compatible",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_key",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007754589896649122,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "numeric_dict_key_alias_compatible",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "lowlevel",
        "kind": "numeric_key",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "accepted",
        "effects": 1,
        "steps": null,
        "scripted_decisions": 0,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "provider_calls": 0
      },
      "result": {
        "status": "accepted",
        "complete": null,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": null,
        "effects": 1,
        "intent_events": 0,
        "scripted_decisions": 0,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_dictionary"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005942089919699356,
      "returned_context_bytes": 56,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "engine_staged_normal",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0012664160021813586,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "engine_staged_normal",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001661166999838315,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "engine_staged_normal",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001458917002310045,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "engine_staged_normal",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013612089969683439,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "engine_staged_normal",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013225830043666065,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "engine_staged_normal",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001282541998079978,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "engine_joint_normal",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 1945,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014426669949898496,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "engine_joint_normal",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 1945,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015258750063367188,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "engine_joint_normal",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 1945,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014812919980613515,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "engine_joint_normal",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 1945,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015640829951735213,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "engine_joint_normal",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 1945,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015850000054342672,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "engine_joint_normal",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "normal",
        "group": "compatibility",
        "strategy": "joint",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 1945,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015985000063665211,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "engine_dependent_fresh_unknown",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dependent",
        "group": "compatibility",
        "unknown_usage": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 2,
        "steps": 2,
        "scripted_decisions": 2,
        "code": null,
        "usage_complete": false,
        "reported_scripted_usage": null,
        "provider_calls": 2
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 2,
        "effects": 2,
        "intent_events": 2,
        "scripted_decisions": 2,
        "provider_calls": 2,
        "received_value_labels": [
          "builtin_integer",
          "builtin_integer"
        ],
        "usage_complete": false,
        "reported_scripted_usage": null,
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2289,
            "shared_context_bytes": 256
          },
          {
            "selection_input_bytes": 2629,
            "shared_context_bytes": 596
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001986083996598609,
      "returned_context_bytes": 793,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "engine_dependent_fresh_unknown",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dependent",
        "group": "compatibility",
        "unknown_usage": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 2,
        "steps": 2,
        "scripted_decisions": 2,
        "code": null,
        "usage_complete": false,
        "reported_scripted_usage": null,
        "provider_calls": 2
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 2,
        "effects": 2,
        "intent_events": 2,
        "scripted_decisions": 2,
        "provider_calls": 2,
        "received_value_labels": [
          "builtin_integer",
          "builtin_integer"
        ],
        "usage_complete": false,
        "reported_scripted_usage": null,
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2289,
            "shared_context_bytes": 256
          },
          {
            "selection_input_bytes": 2629,
            "shared_context_bytes": 596
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.002370708010857925,
      "returned_context_bytes": 793,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "engine_dependent_fresh_unknown",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dependent",
        "group": "compatibility",
        "unknown_usage": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 2,
        "steps": 2,
        "scripted_decisions": 2,
        "code": null,
        "usage_complete": false,
        "reported_scripted_usage": null,
        "provider_calls": 2
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 2,
        "effects": 2,
        "intent_events": 2,
        "scripted_decisions": 2,
        "provider_calls": 2,
        "received_value_labels": [
          "builtin_integer",
          "builtin_integer"
        ],
        "usage_complete": false,
        "reported_scripted_usage": null,
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2289,
            "shared_context_bytes": 256
          },
          {
            "selection_input_bytes": 2629,
            "shared_context_bytes": 596
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.002041749990894459,
      "returned_context_bytes": 793,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "engine_dependent_fresh_unknown",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dependent",
        "group": "compatibility",
        "unknown_usage": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 2,
        "steps": 2,
        "scripted_decisions": 2,
        "code": null,
        "usage_complete": false,
        "reported_scripted_usage": null,
        "provider_calls": 2
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 2,
        "effects": 2,
        "intent_events": 2,
        "scripted_decisions": 2,
        "provider_calls": 2,
        "received_value_labels": [
          "builtin_integer",
          "builtin_integer"
        ],
        "usage_complete": false,
        "reported_scripted_usage": null,
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2289,
            "shared_context_bytes": 256
          },
          {
            "selection_input_bytes": 2629,
            "shared_context_bytes": 596
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017813750018831342,
      "returned_context_bytes": 793,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "engine_dependent_fresh_unknown",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dependent",
        "group": "compatibility",
        "unknown_usage": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 2,
        "steps": 2,
        "scripted_decisions": 2,
        "code": null,
        "usage_complete": false,
        "reported_scripted_usage": null,
        "provider_calls": 2
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 2,
        "effects": 2,
        "intent_events": 2,
        "scripted_decisions": 2,
        "provider_calls": 2,
        "received_value_labels": [
          "builtin_integer",
          "builtin_integer"
        ],
        "usage_complete": false,
        "reported_scripted_usage": null,
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2289,
            "shared_context_bytes": 256
          },
          {
            "selection_input_bytes": 2629,
            "shared_context_bytes": 596
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.002091750007821247,
      "returned_context_bytes": 793,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "engine_dependent_fresh_unknown",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dependent",
        "group": "compatibility",
        "unknown_usage": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 2,
        "steps": 2,
        "scripted_decisions": 2,
        "code": null,
        "usage_complete": false,
        "reported_scripted_usage": null,
        "provider_calls": 2
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 2,
        "effects": 2,
        "intent_events": 2,
        "scripted_decisions": 2,
        "provider_calls": 2,
        "received_value_labels": [
          "builtin_integer",
          "builtin_integer"
        ],
        "usage_complete": false,
        "reported_scripted_usage": null,
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2289,
            "shared_context_bytes": 256
          },
          {
            "selection_input_bytes": 2629,
            "shared_context_bytes": 596
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0020913749904138967,
      "returned_context_bytes": 793,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "unmodified_int_enum_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "int_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "int_enum"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013888749963371083,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "unmodified_int_enum_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "int_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015031669900054112,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "unmodified_int_enum_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "int_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014865829871268943,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "unmodified_int_enum_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "int_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "int_enum"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016927079996094108,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "unmodified_int_enum_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "int_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "int_enum"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001422125002136454,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "unmodified_int_enum_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "int_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014778750046389177,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "unmodified_string_enum_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "string_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "string_enum"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016794169932836667,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "unmodified_string_enum_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "string_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001582124998094514,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "unmodified_string_enum_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "string_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014570000057574362,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "unmodified_string_enum_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "string_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "string_enum"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001247499996679835,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "unmodified_string_enum_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "string_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "string_enum"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016520839999429882,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "unmodified_string_enum_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "string_enum",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001546334009617567,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "unmodified_dict_subclass_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dict_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "dict_subclass"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014647500065620989,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "unmodified_dict_subclass_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dict_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015807500021765009,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "unmodified_dict_subclass_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dict_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014827080012764782,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "unmodified_dict_subclass_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dict_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "dict_subclass"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015789580065757036,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "unmodified_dict_subclass_value",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dict_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "dict_subclass"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013872080016881227,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "unmodified_dict_subclass_value",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "dict_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "execution_uncertain",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "execution_uncertain",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": true,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013983749959152192,
      "returned_context_bytes": 719,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "bind_dict_subclass_metadata",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "metadata_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.00165033400116954,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "bind_dict_subclass_metadata",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "metadata_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "failed",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "failed",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001390542005538009,
      "returned_context_bytes": 675,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "bind_dict_subclass_metadata",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "metadata_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "failed",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "failed",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015375829971162602,
      "returned_context_bytes": 675,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "bind_dict_subclass_metadata",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "metadata_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001368583005387336,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "bind_dict_subclass_metadata",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "metadata_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015836249949643388,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "bind_dict_subclass_metadata",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "metadata_subclass",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "failed",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "failed",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0011157080007251352,
      "returned_context_bytes": 675,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "unmodified_string_enum_tool_id",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "enum_id",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015808750031283125,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "unmodified_string_enum_tool_id",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "enum_id",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "failed",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "failed",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0012568749953061342,
      "returned_context_bytes": 675,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "unmodified_string_enum_tool_id",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "enum_id",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "failed",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "failed",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014032089966349304,
      "returned_context_bytes": 675,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "unmodified_string_enum_tool_id",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "enum_id",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001663625007495284,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "unmodified_string_enum_tool_id",
      "runtime_version": "0.1.16",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "enum_id",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "steps": 1,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "done",
        "complete": true,
        "reason": null,
        "code": null,
        "error_type": null,
        "stopped_guard_phase": null,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [
          "builtin_boolean"
        ],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017279159947065637,
      "returned_context_bytes": 644,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "unmodified_string_enum_tool_id",
      "runtime_version": "0.1.17",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "mode": "engine",
        "kind": "enum_id",
        "group": "compatibility_narrowing",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "failed",
        "effects": 0,
        "steps": 0,
        "scripted_decisions": 1,
        "code": null,
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "provider_calls": 0
      },
      "result": {
        "status": "failed",
        "complete": false,
        "reason": null,
        "code": null,
        "error_type": "ValueError",
        "stopped_guard_phase": null,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "scripted_decisions": 1,
        "provider_calls": 0,
        "received_value_labels": [],
        "usage_complete": true,
        "reported_scripted_usage": {
          "input_tokens": 7,
          "output_tokens": 2
        },
        "action_id_present": false,
        "custom_eq_calls": 0,
        "class_property_reads": 0,
        "request_sizes": [
          {
            "selection_input_bytes": 2201,
            "shared_context_bytes": 256
          }
        ],
        "native_inference_attempts": 0,
        "real_network_attempts": 0,
        "raw_private_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0012499169970396906,
      "returned_context_bytes": 675,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    }
  ]
}