{
  "mode": "offline-synthetic-owned-transport",
  "baseline_commit": "95276bd77eef162645b45bfcd8c40c7a06dac212",
  "candidate": "working tree at invocation; published with its source commit",
  "baseline_source_sha256": {
    "src/apixly_jev_harness/__init__.py": "7093e9f7a45a9b6c75b2cdd3c1c0a2364d1b66afbee46fa5340660e576d52cd6",
    "src/apixly_jev_harness/backend.py": "0ddfab6b270d8219a66614630e70aa10a74cb531f8075d07c255df145725ed5f",
    "src/apixly_jev_harness/callbacks.py": "2d27ad4540d910354c54de7566546b4c9d99b5a5fa87705a1905885349078497",
    "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": "ad5e27cd3eeb8ccf3beb3f6a06e57321edd0fb8c2db63eaa432d3d00ebddf19f",
    "src/apixly_jev_harness/output.py": "94be8b482e4279e29aca128a35caf54744dbe8db8188f14905488f8ddde93a0b",
    "src/apixly_jev_harness/surfaces.py": "65ce73edfe9f6a3c16de373b3a1b85671c3d2cc661f3a50dd906a15721c7d8f1",
    "src/apixly_jev_harness/task.py": "0f5801e9fdb5626e1db65d269c715e8bcf692ebaa3a91f46b9945158b4bbb6db",
    "src/apixly_jev_harness/task_cli.py": "21c883722768c781a68e31e3a602e43eb2a5fac838ea61de478348e9cee87aa8",
    "src/apixly_jev_harness/verify.py": "97d2a178902291094f4e16db08732748d234b8c5ad8dd49a55607c8d195f1ed2"
  },
  "candidate_source_sha256": {
    "src/apixly_jev_harness/__init__.py": "2113d88ead63027f0387c6d3e3292428b6ecb9c80d1bcc47fbcf1bf5882c97fa",
    "src/apixly_jev_harness/backend.py": "f7709b468d3d7c131fb8be42b882e5a0387db0077d1fc07e420153b0447f6754",
    "src/apixly_jev_harness/callbacks.py": "2d27ad4540d910354c54de7566546b4c9d99b5a5fa87705a1905885349078497",
    "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": "488efddf4ade7ba4ab47f959b94c607a40046f7f173b7fe7ac6e5826e38ba049",
    "src/apixly_jev_harness/output.py": "94be8b482e4279e29aca128a35caf54744dbe8db8188f14905488f8ddde93a0b",
    "src/apixly_jev_harness/surfaces.py": "65ce73edfe9f6a3c16de373b3a1b85671c3d2cc661f3a50dd906a15721c7d8f1",
    "src/apixly_jev_harness/task.py": "0f5801e9fdb5626e1db65d269c715e8bcf692ebaa3a91f46b9945158b4bbb6db",
    "src/apixly_jev_harness/task_cli.py": "43eccc327b3d0d1550985b8fcd59e61432a4dc34c5507074d9c43fb8cd876e7c",
    "src/apixly_jev_harness/verify.py": "97d2a178902291094f4e16db08732748d234b8c5ad8dd49a55607c8d195f1ed2"
  },
  "runner_sha256": "139eff2de74bbb60ddef60fcf1630efc9c17ba73ececf84232591df71c03b47b",
  "python_version": "3.14.2",
  "cases": {
    "success": {
      "response": "success"
    },
    "tool_parameter_retry": {
      "response": "parameter_retry",
      "parameter": true
    },
    "three_503_exhausted": {
      "response": "three_503"
    },
    "credential_local_reject": {
      "response": "credential_reject"
    },
    "mock_http_413": {
      "response": "http_413"
    },
    "invalid_response": {
      "response": "invalid"
    },
    "planning_narrow": {
      "response": "success",
      "oversize_description_repetitions": 60000
    },
    "already_done": {
      "response": "success",
      "initially_done": true
    },
    "preused_seed_counters": {
      "response": "success",
      "seed_calls": 1
    },
    "post_effect_string_counter": {
      "response": "success",
      "corrupt_key": "requests",
      "corrupt_kind": "string"
    },
    "post_effect_negative_counter": {
      "response": "success",
      "corrupt_key": "requests",
      "corrupt_kind": "negative"
    },
    "post_effect_bool_counter": {
      "response": "success",
      "corrupt_key": "clients_created",
      "corrupt_kind": "bool"
    },
    "post_effect_private_extra": {
      "response": "success",
      "corrupt_kind": "private_extra"
    },
    "post_effect_reset_to_zero": {
      "response": "success",
      "corrupt_key": "requests",
      "corrupt_kind": "reset_zero"
    },
    "no_active_session": {
      "response": "scripted",
      "without_owned_session": true
    },
    "sdk_encoding_counter_boundary": {
      "response": "surrogate",
      "standalone_sdk": true
    }
  },
  "behavior_parity": true,
  "reproduction": "python benchmarks/owned_transport.py --output local-results/transport-counters-ab.json",
  "scope": "Sixteen fixed synthetic fixtures, three paired repeats with alternating arm order. Normal scenarios use the pinned typed batch and Client only through explicit httpx.MockTransport with a synthetic key; real HTTPTransport is forbidden. Fourteen run-owned snapshots compare fixed SDK counter deltas taken before the run and at terminal finish; valid per-key counters survive while invalid string/negative/bool counters and a reset below a witnessed per-call high-water become unknown, and private extra keys are omitted. A preused Client seed operation is excluded from run-owned deltas. An unwrapped no-active loop uses scripted selection and must omit ownership metadata. A standalone SDK/Session supplement deliberately fails UTF-8 serialization: SDK request counter increments before MockTransport entry, so this is not an HTTP-dispatch count. Actual model/network/UI calls are zero if the guard is never hit.",
  "negative_results": "New fixed snapshot metadata adds returned bytes and local stats-validation work. SDK counters do not prove remote receipt, inference, billing, HTTP dispatch, TCP connection reuse or other tool traffic. Existing telemetry.requests remains logical packed-request aggregation and per-call transport fields remain unchanged. Mocked retry/HTTP failure usage remains unknown; complete local counters do not make it known. Inactive Session and standalone non-loop paths omit metadata rather than report fake zero ownership. No semantic quality, real-provider transport measurement, speed or fee gain is claimed.",
  "cost_conversion": "No tariff or savings assumed. Actual model usage is zero only with no forbidden real transport attempts; a guard hit records unknown actual usage and fails the fixture. Logical requests, scripted selection, mock transport entries, standalone SDK counters and mock token usage are explicitly separate. Real perf_counter timing includes loop admission, pinned request planning, bounded mock retry/validation, effects, terminal snapshot and private journal flush/fsync; excludes process/import/client/task setup, optional seed operation and post-call analysis/source hashing. Loop-only frozen monotonic values stabilize context budgets without substituting for whole-operation timing. Returned bytes exclude local archive paths and scalar context/request sizes do not count provider tokens.",
  "summary": {
    "baseline": {
      "runs": 48,
      "correct": 6,
      "behavior_correct": 48,
      "owned_transport_correct": 6,
      "owned_transport_present": 0,
      "incomplete_owned_counter_rows": 0,
      "logical_requests_reported": 42,
      "scripted_selection_calls": 3,
      "mock_typed_batch_calls": 39,
      "mock_transport_entries_total": 48,
      "mock_seed_transport_entries": 3,
      "mock_transport_entries_in_operation": 45,
      "effects": 27,
      "usage_unknown_rows": 15,
      "real_network_attempts": 0,
      "raw_marker_exposures": 0,
      "median_whole_operation_seconds": 0.001659833505982533,
      "returned_context_bytes": 45627,
      "shared_context_bytes": 12141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      }
    },
    "candidate": {
      "runs": 48,
      "correct": 48,
      "behavior_correct": 48,
      "owned_transport_correct": 48,
      "owned_transport_present": 42,
      "incomplete_owned_counter_rows": 12,
      "logical_requests_reported": 42,
      "scripted_selection_calls": 3,
      "mock_typed_batch_calls": 39,
      "mock_transport_entries_total": 48,
      "mock_seed_transport_entries": 3,
      "mock_transport_entries_in_operation": 45,
      "effects": 27,
      "usage_unknown_rows": 15,
      "real_network_attempts": 0,
      "raw_marker_exposures": 0,
      "median_whole_operation_seconds": 0.0016981250009848736,
      "returned_context_bytes": 53319,
      "shared_context_bytes": 12141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      }
    }
  },
  "rows": [
    {
      "case": "success",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0013451669947244227,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "success",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016678749962011352,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "success",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001772207993781194,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "success",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0016602090036030859,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "success",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.001782124993042089,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "success",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001592957996763289,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "tool_parameter_retry",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "parameter_retry",
        "parameter": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 2,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 2,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 2,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          },
          {
            "stage": "parameter_selection",
            "shared_context_bytes": 575,
            "shared_context_characters": 575,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 2,
              "clients_created": 0,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2619,
            "json_utf8_bytes": 2619,
            "stage": "tool_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.002461458003381267,
      "returned_context_bytes": 1889,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "tool_parameter_retry",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "parameter_retry",
        "parameter": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 2,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 2,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 2,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          },
          {
            "stage": "parameter_selection",
            "shared_context_bytes": 575,
            "shared_context_characters": 575,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 2,
              "clients_created": 0,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2619,
            "json_utf8_bytes": 2619,
            "stage": "tool_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.002682417005416937,
      "returned_context_bytes": 2071,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "tool_parameter_retry",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "parameter_retry",
        "parameter": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 2,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 2,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 2,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          },
          {
            "stage": "parameter_selection",
            "shared_context_bytes": 575,
            "shared_context_characters": 575,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 2,
              "clients_created": 0,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2619,
            "json_utf8_bytes": 2619,
            "stage": "tool_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0024233340081991628,
      "returned_context_bytes": 2071,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "tool_parameter_retry",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "parameter_retry",
        "parameter": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 2,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 2,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 2,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          },
          {
            "stage": "parameter_selection",
            "shared_context_bytes": 575,
            "shared_context_characters": 575,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 2,
              "clients_created": 0,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2619,
            "json_utf8_bytes": 2619,
            "stage": "tool_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0025194170011673123,
      "returned_context_bytes": 1889,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "tool_parameter_retry",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "parameter_retry",
        "parameter": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 2,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 2,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 2,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          },
          {
            "stage": "parameter_selection",
            "shared_context_bytes": 575,
            "shared_context_characters": 575,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 2,
              "clients_created": 0,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2619,
            "json_utf8_bytes": 2619,
            "stage": "tool_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.002238457993371412,
      "returned_context_bytes": 1889,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "tool_parameter_retry",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "parameter_retry",
        "parameter": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 2,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 2,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 2,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          },
          {
            "stage": "parameter_selection",
            "shared_context_bytes": 575,
            "shared_context_characters": 575,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 2,
              "clients_created": 0,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2619,
            "json_utf8_bytes": 2619,
            "stage": "tool_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          },
          {
            "json_characters": 1590,
            "json_utf8_bytes": 1590,
            "stage": "parameter_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0022326670004986227,
      "returned_context_bytes": 2071,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "three_503_exhausted",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "three_503",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 3,
              "clients_created": 1,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0017564999870955944,
      "returned_context_bytes": 1072,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "three_503_exhausted",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "three_503",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 3,
              "clients_created": 1,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017625000036787242,
      "returned_context_bytes": 1254,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "three_503_exhausted",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "three_503",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 3,
              "clients_created": 1,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0020237089920556173,
      "returned_context_bytes": 1254,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "three_503_exhausted",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "three_503",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 3,
              "clients_created": 1,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.001909541999339126,
      "returned_context_bytes": 1072,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "three_503_exhausted",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "three_503",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 3,
              "clients_created": 1,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0018824580038199201,
      "returned_context_bytes": 1072,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "three_503_exhausted",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "three_503",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 3,
        "sdk_attempt_counter_in_operation": 3,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 3,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 3,
        "sdk_counter_delta_independent": {
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 3,
          "clients_created": 1,
          "client_reuses": 2,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 3,
              "clients_created": 1,
              "client_reuses": 2
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0018614590080687776,
      "returned_context_bytes": 1254,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "credential_local_reject",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "credential_reject",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 0,
              "clients_created": 0,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0011288749956293032,
      "returned_context_bytes": 1062,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "credential_local_reject",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "credential_reject",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 0,
              "clients_created": 0,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0011714169959304854,
      "returned_context_bytes": 1244,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "credential_local_reject",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "credential_reject",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 0,
              "clients_created": 0,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0012611660058610141,
      "returned_context_bytes": 1244,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "credential_local_reject",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "credential_reject",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 0,
              "clients_created": 0,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0012092079996364191,
      "returned_context_bytes": 1062,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "credential_local_reject",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "credential_reject",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 0,
              "clients_created": 0,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0011441669921623543,
      "returned_context_bytes": 1062,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "credential_local_reject",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "credential_reject",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 0,
              "clients_created": 0,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001335082997684367,
      "returned_context_bytes": 1244,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "mock_http_413",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "http_413",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0016478750039823353,
      "returned_context_bytes": 1072,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "mock_http_413",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "http_413",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016196249925997108,
      "returned_context_bytes": 1254,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "mock_http_413",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "http_413",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001520000005257316,
      "returned_context_bytes": 1254,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "mock_http_413",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "http_413",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0016066660027718171,
      "returned_context_bytes": 1072,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "mock_http_413",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "http_413",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0014014589978614822,
      "returned_context_bytes": 1072,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "mock_http_413",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "http_413",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014047919976292178,
      "returned_context_bytes": 1254,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "invalid_response",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "invalid",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0016069580015027896,
      "returned_context_bytes": 1084,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "invalid_response",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "invalid",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015359999961219728,
      "returned_context_bytes": 1266,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "invalid_response",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "invalid",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016778749995864928,
      "returned_context_bytes": 1266,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "invalid_response",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "invalid",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0014024590054759756,
      "returned_context_bytes": 1084,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "invalid_response",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "invalid",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.00299950000771787,
      "returned_context_bytes": 1084,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "invalid_response",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "invalid",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": false,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": false,
        "reported_mock_usage": null,
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": false,
            "usage_known": false,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001476249992265366,
      "returned_context_bytes": 1266,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "planning_narrow",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "oversize_description_repetitions": 60000,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 0,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0015565419889753684,
      "returned_context_bytes": 804,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "planning_narrow",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "oversize_description_repetitions": 60000,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 0,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.002240500005427748,
      "returned_context_bytes": 986,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "planning_narrow",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "oversize_description_repetitions": 60000,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 0,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001352792009129189,
      "returned_context_bytes": 986,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "planning_narrow",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "oversize_description_repetitions": 60000,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 0,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.001761708001140505,
      "returned_context_bytes": 804,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "planning_narrow",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "oversize_description_repetitions": 60000,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 0,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0016594580083619803,
      "returned_context_bytes": 804,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "planning_narrow",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "oversize_description_repetitions": 60000,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "needs_review",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "needs_review",
        "exception_type": null,
        "complete": false,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 0,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017241669993381947,
      "returned_context_bytes": 986,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "already_done",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "initially_done": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0008577920089010149,
      "returned_context_bytes": 213,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "already_done",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "initially_done": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008039589883992448,
      "returned_context_bytes": 395,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "already_done",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "initially_done": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008517920068698004,
      "returned_context_bytes": 395,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "already_done",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "initially_done": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.000873249999131076,
      "returned_context_bytes": 213,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "already_done",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "initially_done": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0008845830016070977,
      "returned_context_bytes": 213,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "already_done",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "initially_done": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 0,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008734580042073503,
      "returned_context_bytes": 395,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "preused_seed_counters",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "seed_calls": 1,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 2,
        "mock_seed_transport_entries": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 0,
              "client_reuses": 1
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 186,
            "json_utf8_bytes": 186,
            "stage": "seed"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0014312080020317808,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "preused_seed_counters",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "seed_calls": 1,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 2,
        "mock_seed_transport_entries": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 0,
              "client_reuses": 1
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 186,
            "json_utf8_bytes": 186,
            "stage": "seed"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001493374991696328,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "preused_seed_counters",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "seed_calls": 1,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 2,
        "mock_seed_transport_entries": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 0,
              "client_reuses": 1
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 186,
            "json_utf8_bytes": 186,
            "stage": "seed"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016372499958379194,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "preused_seed_counters",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "seed_calls": 1,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 2,
        "mock_seed_transport_entries": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 0,
              "client_reuses": 1
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 186,
            "json_utf8_bytes": 186,
            "stage": "seed"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0015491660014959052,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "preused_seed_counters",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "seed_calls": 1,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 2,
        "mock_seed_transport_entries": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 0,
              "client_reuses": 1
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 186,
            "json_utf8_bytes": 186,
            "stage": "seed"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0015170410042628646,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "preused_seed_counters",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "seed_calls": 1,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 2,
        "mock_seed_transport_entries": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 0,
          "client_reuses": 1,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 0,
              "client_reuses": 1
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 186,
            "json_utf8_bytes": 186,
            "stage": "seed"
          },
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017034580087056383,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "post_effect_string_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "string",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "str",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0018559590098448098,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "post_effect_string_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "string",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "str",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0018148340022889897,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "post_effect_string_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "string",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "str",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017579160048626363,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "post_effect_string_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "string",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "str",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0019094589952146634,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "post_effect_string_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "string",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "str",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.001923417003126815,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "post_effect_string_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "string",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "str",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0019606249989010394,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "post_effect_negative_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "negative",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0019323749875184149,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "post_effect_negative_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "negative",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0018623750074766576,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "post_effect_negative_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "negative",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0019253329955972731,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "post_effect_negative_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "negative",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0017282500048168004,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "post_effect_negative_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "negative",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0017681250028545037,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "post_effect_negative_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "negative",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017849170108092949,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "post_effect_bool_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "clients_created",
        "corrupt_kind": "bool",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "bool",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0019089999987045303,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "post_effect_bool_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "clients_created",
        "corrupt_kind": "bool",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "bool",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.002122249992680736,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "post_effect_bool_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "clients_created",
        "corrupt_kind": "bool",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "bool",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0018680830107768998,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "post_effect_bool_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "clients_created",
        "corrupt_kind": "bool",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "bool",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0014202909951563925,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "post_effect_bool_counter",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "clients_created",
        "corrupt_kind": "bool",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "bool",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0018520419980632141,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "post_effect_bool_counter",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "clients_created",
        "corrupt_kind": "bool",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "bool",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": null,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0019317499973112717,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "post_effect_private_extra",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_kind": "private_extra",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0018022089934675023,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "post_effect_private_extra",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_kind": "private_extra",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0018482920131646097,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "post_effect_private_extra",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_kind": "private_extra",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001692791993264109,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "post_effect_private_extra",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_kind": "private_extra",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0019177919893991202,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "post_effect_private_extra",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_kind": "private_extra",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.002116208997904323,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "post_effect_private_extra",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_kind": "private_extra",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": true
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0019937920005759224,
      "returned_context_bytes": 1255,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "post_effect_reset_to_zero",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "reset_zero",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0017715419962769374,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "post_effect_reset_to_zero",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "reset_zero",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0017297919985139742,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "post_effect_reset_to_zero",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "reset_zero",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0019389579974813387,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "post_effect_reset_to_zero",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "reset_zero",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.001959000001079403,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "post_effect_reset_to_zero",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "reset_zero",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": false,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0024613330024294555,
      "returned_context_bytes": 1073,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "post_effect_reset_to_zero",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "success",
        "corrupt_key": "requests",
        "corrupt_kind": "reset_zero",
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 1,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": true,
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        }
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 1,
        "mock_transport_entries_total": 1,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 1,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": {
          "scope": "run_owned_provider_client",
          "snapshot": "terminal_finish",
          "request_attempts": null,
          "clients_created": 1,
          "client_reuses": 0,
          "counters_complete": false
        },
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 5,
          "output_tokens": 1
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": "jev-synthetic-resolved",
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": {
              "requests": 1,
              "clients_created": 1,
              "client_reuses": 0
            }
          }
        ],
        "mock_request_sizes": [
          {
            "json_characters": 2237,
            "json_utf8_bytes": 2237,
            "stage": "tool_selection"
          }
        ],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001850625005317852,
      "returned_context_bytes": 1259,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "no_active_session",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "scripted",
        "without_owned_session": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": null
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 1,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001369041987345554,
      "returned_context_bytes": 462,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "no_active_session",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "scripted",
        "without_owned_session": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": null
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 1,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013512500008800998,
      "returned_context_bytes": 462,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "no_active_session",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "scripted",
        "without_owned_session": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": null
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 1,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0010507090046303347,
      "returned_context_bytes": 462,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "no_active_session",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "scripted",
        "without_owned_session": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": null
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 1,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015748340083518997,
      "returned_context_bytes": 462,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "no_active_session",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "scripted",
        "without_owned_session": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": null
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 1,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013752090017078444,
      "returned_context_bytes": 462,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "no_active_session",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "scripted",
        "without_owned_session": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "done",
        "effects": 1,
        "logical_requests": 1,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 0,
        "usage_complete": true,
        "owned_transport": null
      },
      "result": {
        "status": "done",
        "exception_type": null,
        "complete": true,
        "steps": 1,
        "effects": 1,
        "intent_events": 1,
        "logical_requests": 1,
        "scripted_selection_calls": 1,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 0,
          "clients_created": 0,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": true,
        "reported_mock_usage": {
          "input_tokens": 0,
          "output_tokens": 0
        },
        "evaluation_calls": [
          {
            "stage": "tool_selection",
            "shared_context_bytes": 248,
            "shared_context_characters": 248,
            "logical_requests_reported": 1,
            "model_resolved": null,
            "usage_complete": true,
            "usage_known": true,
            "per_call_transport": null
          }
        ],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.001395333994878456,
      "returned_context_bytes": 462,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "sdk_encoding_counter_boundary",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "surrogate",
        "standalone_sdk": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "sdk_encoding_error",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": null,
        "owned_transport": null
      },
      "result": {
        "status": "sdk_encoding_error",
        "exception_type": "UnicodeEncodeError",
        "complete": null,
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": null,
        "reported_mock_usage": null,
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0004581669927574694,
      "returned_context_bytes": 40,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "sdk_encoding_counter_boundary",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "surrogate",
        "standalone_sdk": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "sdk_encoding_error",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": null,
        "owned_transport": null
      },
      "result": {
        "status": "sdk_encoding_error",
        "exception_type": "UnicodeEncodeError",
        "complete": null,
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": null,
        "reported_mock_usage": null,
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006143749924376607,
      "returned_context_bytes": 40,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "sdk_encoding_counter_boundary",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "surrogate",
        "standalone_sdk": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "sdk_encoding_error",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": null,
        "owned_transport": null
      },
      "result": {
        "status": "sdk_encoding_error",
        "exception_type": "UnicodeEncodeError",
        "complete": null,
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": null,
        "reported_mock_usage": null,
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006367499881889671,
      "returned_context_bytes": 40,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "sdk_encoding_counter_boundary",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "surrogate",
        "standalone_sdk": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "sdk_encoding_error",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": null,
        "owned_transport": null
      },
      "result": {
        "status": "sdk_encoding_error",
        "exception_type": "UnicodeEncodeError",
        "complete": null,
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": null,
        "reported_mock_usage": null,
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006190409912960604,
      "returned_context_bytes": 40,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "sdk_encoding_counter_boundary",
      "runtime_version": "0.1.11",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "surrogate",
        "standalone_sdk": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "sdk_encoding_error",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": null,
        "owned_transport": null
      },
      "result": {
        "status": "sdk_encoding_error",
        "exception_type": "UnicodeEncodeError",
        "complete": null,
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": null,
        "reported_mock_usage": null,
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005841250094817951,
      "returned_context_bytes": 40,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "sdk_encoding_counter_boundary",
      "runtime_version": "0.1.12",
      "runtime_loaded_from_requested_source": true,
      "fixture": {
        "response": "surrogate",
        "standalone_sdk": true,
        "frozen_loop_monotonic": 100.0
      },
      "expected": {
        "status": "sdk_encoding_error",
        "effects": 0,
        "logical_requests": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_attempt_counter_in_operation": 1,
        "usage_complete": null,
        "owned_transport": null
      },
      "result": {
        "status": "sdk_encoding_error",
        "exception_type": "UnicodeEncodeError",
        "complete": null,
        "steps": null,
        "effects": 0,
        "intent_events": 0,
        "logical_requests": 0,
        "scripted_selection_calls": 0,
        "mock_typed_batch_calls": 0,
        "mock_transport_entries_total": 0,
        "mock_seed_transport_entries": 0,
        "mock_transport_entries_in_operation": 0,
        "sdk_counter_delta_independent": {
          "request_attempts": 1,
          "clients_created": 1,
          "client_reuses": 0
        },
        "counter_types_at_finish": {
          "request_attempts": "int",
          "clients_created": "int",
          "client_reuses": "int"
        },
        "owned_transport": null,
        "owned_transport_correct": true,
        "behavior_correct": true,
        "usage_complete": null,
        "reported_mock_usage": null,
        "evaluation_calls": [],
        "mock_request_sizes": [],
        "real_network_attempts": 0,
        "raw_marker_exposures": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006184999947436154,
      "returned_context_bytes": 40,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    }
  ]
}