{
  "mode": "offline-synthetic-cli-arguments",
  "baseline_commit": "609d76822dc721a1f19bb7c3b2837f553507d24f",
  "candidate": "working tree at invocation; published with its source commit",
  "baseline_source_sha256": {
    "src/apixly_jev_harness/__init__.py": "2e0428c6d8bbc5f74528d1fa503609a42145281340822f20bdd6b2d1f49b4b99",
    "src/apixly_jev_harness/backend.py": "e4b4c80d88c63648c1392f41eb8796430fcbce6f81d1061c71519d0759432eda",
    "src/apixly_jev_harness/callbacks.py": "2d27ad4540d910354c54de7566546b4c9d99b5a5fa87705a1905885349078497",
    "src/apixly_jev_harness/cli.py": "8d9434bf8a8b539163194a84e3f099761d08ef32201f99cbce46c139f3726d94",
    "src/apixly_jev_harness/context.py": "e0deafa0addd0f96b422a497dae8176d059dd36635954d6770c0cf61ba8c3593",
    "src/apixly_jev_harness/diagnostics.py": "24950b4d834d9e6c902996c7bcddfa208944e42300ae062e9ebc2a44f75a091f",
    "src/apixly_jev_harness/engine.py": "cdedf1ef3a769cc1557dfa63cf6b6e6c4b48a14eb7b0fc12d3d9a957a53b3be2",
    "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": "1a257875aa9643c7330ab4484763209efd2e3acd1d883c26b2604f7b8bbbfe9c",
    "src/apixly_jev_harness/verify.py": "97d2a178902291094f4e16db08732748d234b8c5ad8dd49a55607c8d195f1ed2"
  },
  "candidate_source_sha256": {
    "src/apixly_jev_harness/__init__.py": "d3701078fca6c5e7ae6c2fffeb1f00a327dae07265b73c94aa99e0a3e16fd4d4",
    "src/apixly_jev_harness/backend.py": "e4b4c80d88c63648c1392f41eb8796430fcbce6f81d1061c71519d0759432eda",
    "src/apixly_jev_harness/callbacks.py": "2d27ad4540d910354c54de7566546b4c9d99b5a5fa87705a1905885349078497",
    "src/apixly_jev_harness/cli.py": "8d9434bf8a8b539163194a84e3f099761d08ef32201f99cbce46c139f3726d94",
    "src/apixly_jev_harness/context.py": "e0deafa0addd0f96b422a497dae8176d059dd36635954d6770c0cf61ba8c3593",
    "src/apixly_jev_harness/diagnostics.py": "24950b4d834d9e6c902996c7bcddfa208944e42300ae062e9ebc2a44f75a091f",
    "src/apixly_jev_harness/engine.py": "cdedf1ef3a769cc1557dfa63cf6b6e6c4b48a14eb7b0fc12d3d9a957a53b3be2",
    "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": "e7ae2e5c9d6c4179997568774769fbcab292c246f8ee3bf8031c61e4f6a03f37",
    "src/apixly_jev_harness/verify.py": "97d2a178902291094f4e16db08732748d234b8c5ad8dd49a55607c8d195f1ed2"
  },
  "runner_sha256": "1c9d5d99080afc75b6752bdbad5f86d430032f0f38429f25d4e60fe0b25a9df6",
  "synthetic_task_sha256": "e5ec6132e5c040d3a63a37c46695960260445e1f69281dda9dc6e85949c35108",
  "python_version": "3.14.2",
  "cases": {
    "missing_command": {
      "argv": [],
      "kind": "arguments",
      "exit_code": 2
    },
    "unknown_command": {
      "argv": [
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "unknown_flag": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "unknown_flag_value": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--unknown",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "invalid_integer": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--max-steps",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "invalid_float": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--timeout",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "missing_goal": {
      "argv": [
        "run",
        "$SCRIPT"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "missing_goal_value": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "extra_positional": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "missing_run_script": {
      "argv": [
        "run",
        "--goal",
        "Observe synthetic completion"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "missing_check_script": {
      "argv": [
        "check-task"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "missing_init_script": {
      "argv": [
        "init-task"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "ambiguous_short_option": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--m",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "missing_inputs_value": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "init_unknown_flag": {
      "argv": [
        "init-task",
        "$NEW",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "spec_extra_value": {
      "argv": [
        "spec",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "arguments",
      "exit_code": 2
    },
    "root_help": {
      "argv": [
        "--help"
      ],
      "kind": "help",
      "exit_code": 0
    },
    "run_help": {
      "argv": [
        "run",
        "--help"
      ],
      "kind": "help",
      "exit_code": 0
    },
    "help_before_unknown_flag": {
      "argv": [
        "run",
        "--help",
        "--unknown"
      ],
      "kind": "help",
      "exit_code": 0
    },
    "valid_spec": {
      "argv": [
        "spec"
      ],
      "kind": "spec",
      "exit_code": 0
    },
    "valid_check": {
      "argv": [
        "check-task",
        "$SCRIPT"
      ],
      "kind": "check",
      "exit_code": 0
    },
    "valid_init": {
      "argv": [
        "init-task",
        "$NEW"
      ],
      "kind": "init",
      "exit_code": 0
    },
    "valid_complete_run": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--archive",
        "$ARCHIVE"
      ],
      "kind": "run",
      "exit_code": 0
    },
    "valid_long_abbreviations": {
      "argv": [
        "run",
        "$SCRIPT",
        "--go",
        "Observe synthetic completion",
        "--max-s",
        "2",
        "--time",
        "5",
        "--archive",
        "$ARCHIVE"
      ],
      "kind": "run",
      "exit_code": 0
    },
    "malformed_json_configuration": {
      "argv": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs",
        "{SYNTHETIC_ARGUMENT_VALUE"
      ],
      "kind": "configuration",
      "exit_code": 2
    }
  },
  "reproduction": "python benchmarks/cli_arguments.py --output local-results/cli-arguments-ab.json",
  "scope": "Twenty-five fixed synthetic argv fixtures, three paired repeats with alternating arm order. Sixteen malformed parser inputs expect a fixed arguments-phase JSON failure and zero task source read/spec/exec, import/init/journal effects and actual model calls. Help remains plain text with SystemExit(0), including help before a trailing unknown flag; valid long-option abbreviations remain accepted, and malformed JSON remains a distinct configuration-phase failure. Valid task runs import only a synthetic local marker and an already-complete workflow, with zero tool effects/decisions. Native inference entrypoints are forbidden. Raw captured stdout/stderr stay local to each trial; the report contains counts/digests/fixed synthetic marker exposures and argv templates with path placeholders.",
  "negative_results": "Fixed failure metadata adds bytes for terse failures and omits argparse's specific corrective error/help detail. AI callers can inspect spec/help but this does not guarantee every malformed input or every process failure returns JSON. Argparse help can exit successfully before validating later unknown arguments; that plain-text precedence is preserved. BaseExceptions, failed output streams, encoding issues, trusted task runtime and sandboxing are outside the argument-parser boundary. Source-access probes cover the synthetic task content/import paths, not every operating-system filesystem read. Valid help is intentionally not a JSON packet. No semantic quality, speed or fee gain is claimed.",
  "cost_conversion": "Zero actual model calls only with no native guard attempts; a guard hit records unknown usage and fails the fixture. No tariff or savings assumed. Whole-operation timing includes parser construction/admission, permitted offline commands or trusted synthetic task import/journaling, output serialization and stream restoration; excludes process/import/temp fixture setup and result analysis. Returned context bytes measure all actual stdout plus stderr, including baseline help/error detail. Fixed synthetic marker exposure counts do not establish arbitrary data confidentiality.",
  "summary": {
    "baseline": {
      "runs": 75,
      "argument_error_runs": 48,
      "help_runs": 9,
      "exact_json_packets": 18,
      "correct": 27,
      "status_metadata_correct": 27,
      "filesystem_policy_correct": 75,
      "source_access_policy_correct": 75,
      "streams_restored": 75,
      "raw_marker_exposures": {
        "stdout": 0,
        "stderr": 24
      },
      "native_inference_attempts": 0,
      "median_whole_operation_seconds": 0.0006324169953586534,
      "returned_context_bytes": 39492,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      }
    },
    "candidate": {
      "runs": 75,
      "argument_error_runs": 48,
      "help_runs": 9,
      "exact_json_packets": 66,
      "correct": 75,
      "status_metadata_correct": 75,
      "filesystem_policy_correct": 75,
      "source_access_policy_correct": 75,
      "streams_restored": 75,
      "raw_marker_exposures": {
        "stdout": 0,
        "stderr": 0
      },
      "native_inference_attempts": 0,
      "median_whole_operation_seconds": 0.0005872079927939922,
      "returned_context_bytes": 34578,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      }
    }
  },
  "rows": [
    {
      "case": "missing_command",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 142,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "ffbacd07cfb1bbf002b47c96a30ba5ef60463812cf8612962fb799fce442d3ba",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006630840071011335,
      "returned_context_bytes": 142,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "missing_command",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005303329962771386,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "missing_command",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005320000054780394,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "missing_command",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 142,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "ffbacd07cfb1bbf002b47c96a30ba5ef60463812cf8612962fb799fce442d3ba",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005505000008270144,
      "returned_context_bytes": 142,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "missing_command",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 142,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "ffbacd07cfb1bbf002b47c96a30ba5ef60463812cf8612962fb799fce442d3ba",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005344580131350085,
      "returned_context_bytes": 142,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "missing_command",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005208749935263768,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "unknown_command",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 211,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "86b43d07bdd2b99ff91648baacf0f221c67ef0ddaf43139569cdc38bb9598516",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007069580024108291,
      "returned_context_bytes": 211,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "unknown_command",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000544916998478584,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "unknown_command",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005270420078886673,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "unknown_command",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 211,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "86b43d07bdd2b99ff91648baacf0f221c67ef0ddaf43139569cdc38bb9598516",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006324169953586534,
      "returned_context_bytes": 211,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "unknown_command",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 211,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "86b43d07bdd2b99ff91648baacf0f221c67ef0ddaf43139569cdc38bb9598516",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.000564083005883731,
      "returned_context_bytes": 211,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "unknown_command",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005427079886430874,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 147,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "364b58c995dca7ea5df12c38e0dee8846fa64aaec98c279f3acc7c4487acc9c8",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006050829979358241,
      "returned_context_bytes": 147,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005865419952897355,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005782080115750432,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 147,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "364b58c995dca7ea5df12c38e0dee8846fa64aaec98c279f3acc7c4487acc9c8",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006167499959701672,
      "returned_context_bytes": 147,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 147,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "364b58c995dca7ea5df12c38e0dee8846fa64aaec98c279f3acc7c4487acc9c8",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005848750006407499,
      "returned_context_bytes": 147,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005840420053573325,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "unknown_flag_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--unknown",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 155,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "1f279a652102ebce20f1c9e4074c2ef2c7818a3a8b3f73065f50c327831ee333",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006063750042812899,
      "returned_context_bytes": 155,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "unknown_flag_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--unknown",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006010829965816811,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "unknown_flag_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--unknown",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006124170031398535,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "unknown_flag_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--unknown",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 155,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "1f279a652102ebce20f1c9e4074c2ef2c7818a3a8b3f73065f50c327831ee333",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005957089888397604,
      "returned_context_bytes": 155,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "unknown_flag_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--unknown",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 155,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "1f279a652102ebce20f1c9e4074c2ef2c7818a3a8b3f73065f50c327831ee333",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006046250055078417,
      "returned_context_bytes": 155,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "unknown_flag_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--unknown",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005782910011475906,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "invalid_integer",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--max-steps",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 451,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "8d8767a1e5741608d0f3baa292e6f3ea0e5fc357f211e2c154981ac69023aaa0",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006164999940665439,
      "returned_context_bytes": 451,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "invalid_integer",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--max-steps",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005839170044055209,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "invalid_integer",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--max-steps",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005853750044479966,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "invalid_integer",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--max-steps",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 451,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "8d8767a1e5741608d0f3baa292e6f3ea0e5fc357f211e2c154981ac69023aaa0",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006297080108197406,
      "returned_context_bytes": 451,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "invalid_integer",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--max-steps",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 451,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "8d8767a1e5741608d0f3baa292e6f3ea0e5fc357f211e2c154981ac69023aaa0",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006438330019591376,
      "returned_context_bytes": 451,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "invalid_integer",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--max-steps",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006112499977461994,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "invalid_float",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--timeout",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 451,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "96b55114354d93e7b286e685cf9914ab26b8b9cb45ea673e02a0e0647949a562",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006329589959932491,
      "returned_context_bytes": 451,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "invalid_float",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--timeout",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005775000026915222,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "invalid_float",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--timeout",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005837079952470958,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "invalid_float",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--timeout",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 451,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "96b55114354d93e7b286e685cf9914ab26b8b9cb45ea673e02a0e0647949a562",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006568329990841448,
      "returned_context_bytes": 451,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "invalid_float",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--timeout",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 451,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "96b55114354d93e7b286e685cf9914ab26b8b9cb45ea673e02a0e0647949a562",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006537080043926835,
      "returned_context_bytes": 451,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "invalid_float",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--timeout",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005819999933009967,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "missing_goal",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 428,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "ddabd9325bf55f4010ffa5e4f6a95cdabea6dddb5923f8648bb532d5165e8fb0",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.000633040996035561,
      "returned_context_bytes": 428,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "missing_goal",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006933749973541126,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "missing_goal",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005614999972749501,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "missing_goal",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 428,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "ddabd9325bf55f4010ffa5e4f6a95cdabea6dddb5923f8648bb532d5165e8fb0",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.000591584001085721,
      "returned_context_bytes": 428,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "missing_goal",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 428,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "ddabd9325bf55f4010ffa5e4f6a95cdabea6dddb5923f8648bb532d5165e8fb0",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007404580101137981,
      "returned_context_bytes": 428,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "missing_goal",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006346669979393482,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "missing_goal_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 422,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "88f683c5f2198c63cd9fe81de15c08b89d1f045b9c46778c0dbc4b4d5fc7fbdf",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006720839883200824,
      "returned_context_bytes": 422,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "missing_goal_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006525839999085292,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "missing_goal_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006629170093219727,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "missing_goal_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 422,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "88f683c5f2198c63cd9fe81de15c08b89d1f045b9c46778c0dbc4b4d5fc7fbdf",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006259579968173057,
      "returned_context_bytes": 422,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "missing_goal_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 422,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "88f683c5f2198c63cd9fe81de15c08b89d1f045b9c46778c0dbc4b4d5fc7fbdf",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006690000009257346,
      "returned_context_bytes": 422,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "missing_goal_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007607500010635704,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "extra_positional",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 145,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "439662062674271da71f42300efa3675bce929cc1eb01890cbac32a3692ecd6e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005947500030742958,
      "returned_context_bytes": 145,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "extra_positional",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005815839977003634,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "extra_positional",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005677909939549863,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "extra_positional",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 145,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "439662062674271da71f42300efa3675bce929cc1eb01890cbac32a3692ecd6e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006035830010659993,
      "returned_context_bytes": 145,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "extra_positional",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 145,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "439662062674271da71f42300efa3675bce929cc1eb01890cbac32a3692ecd6e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005895419890293851,
      "returned_context_bytes": 145,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "extra_positional",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005759169871453196,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "missing_run_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--goal",
        "Observe synthetic completion"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 428,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "dc5a1223f9c8d13507ed1b39cf97c2e42686b0a3d5a00de03125d2e6cd55404e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006199159979587421,
      "returned_context_bytes": 428,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "missing_run_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--goal",
        "Observe synthetic completion"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006818330002715811,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "missing_run_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--goal",
        "Observe synthetic completion"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007329170039156452,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "missing_run_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--goal",
        "Observe synthetic completion"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 428,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "dc5a1223f9c8d13507ed1b39cf97c2e42686b0a3d5a00de03125d2e6cd55404e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006192080036271363,
      "returned_context_bytes": 428,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "missing_run_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--goal",
        "Observe synthetic completion"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 428,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "dc5a1223f9c8d13507ed1b39cf97c2e42686b0a3d5a00de03125d2e6cd55404e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0007825000066077337,
      "returned_context_bytes": 428,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "missing_run_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--goal",
        "Observe synthetic completion"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005820829974254593,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "missing_check_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 128,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "c3cf2c0e295c63135a86ce0c2063ceff7bb5baa4997812980000171692e6bbf9",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005750420095864683,
      "returned_context_bytes": 128,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "missing_check_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005557500116992742,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "missing_check_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005668750090990216,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "missing_check_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 128,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "c3cf2c0e295c63135a86ce0c2063ceff7bb5baa4997812980000171692e6bbf9",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005891250038985163,
      "returned_context_bytes": 128,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "missing_check_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 128,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "c3cf2c0e295c63135a86ce0c2063ceff7bb5baa4997812980000171692e6bbf9",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006336660007946193,
      "returned_context_bytes": 128,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "missing_check_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005532080103876069,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "missing_init_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 126,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "98eaaa0a28bf2c995059079043ed6031cbaaac576fadb9986f77a58e028ab0a5",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005637920112349093,
      "returned_context_bytes": 126,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "missing_init_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005640000017592683,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "missing_init_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005625829944619909,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "missing_init_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 126,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "98eaaa0a28bf2c995059079043ed6031cbaaac576fadb9986f77a58e028ab0a5",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005694580031558871,
      "returned_context_bytes": 126,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "missing_init_script",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 126,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "98eaaa0a28bf2c995059079043ed6031cbaaac576fadb9986f77a58e028ab0a5",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005817089986521751,
      "returned_context_bytes": 126,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "missing_init_script",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005671660037478432,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "ambiguous_short_option",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--m",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 441,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "30629cc5e317ccbb13d276f563b0023e7f3f585e35415c5a73474393e88f8121",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006204160017659888,
      "returned_context_bytes": 441,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "ambiguous_short_option",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--m",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005976249958621338,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "ambiguous_short_option",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--m",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005867500003660098,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "ambiguous_short_option",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--m",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 441,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "30629cc5e317ccbb13d276f563b0023e7f3f585e35415c5a73474393e88f8121",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006434999959310517,
      "returned_context_bytes": 441,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "ambiguous_short_option",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--m",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 441,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "30629cc5e317ccbb13d276f563b0023e7f3f585e35415c5a73474393e88f8121",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006434590031858534,
      "returned_context_bytes": 441,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "ambiguous_short_option",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--m",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005678749876096845,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "missing_inputs_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 424,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "234d2a6b8c9b87c7f70554334feab86e9f71d756d108a4eeaec28d9155f00531",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006347909948090091,
      "returned_context_bytes": 424,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "missing_inputs_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005918750102864578,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "missing_inputs_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005872079927939922,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "missing_inputs_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 424,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "234d2a6b8c9b87c7f70554334feab86e9f71d756d108a4eeaec28d9155f00531",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006369580078171566,
      "returned_context_bytes": 424,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "missing_inputs_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 424,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "234d2a6b8c9b87c7f70554334feab86e9f71d756d108a4eeaec28d9155f00531",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0006224169919732958,
      "returned_context_bytes": 424,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "missing_inputs_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006180839991429821,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "init_unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 147,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "364b58c995dca7ea5df12c38e0dee8846fa64aaec98c279f3acc7c4487acc9c8",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.000592875003349036,
      "returned_context_bytes": 147,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "init_unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005831250018673018,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "init_unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005684999923687428,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "init_unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 147,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "364b58c995dca7ea5df12c38e0dee8846fa64aaec98c279f3acc7c4487acc9c8",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005960409907856956,
      "returned_context_bytes": 147,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "init_unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 147,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "364b58c995dca7ea5df12c38e0dee8846fa64aaec98c279f3acc7c4487acc9c8",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005926250014454126,
      "returned_context_bytes": 147,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "init_unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW",
        "--SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006462089950218797,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "spec_extra_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 145,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "439662062674271da71f42300efa3675bce929cc1eb01890cbac32a3692ecd6e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005708749959012493,
      "returned_context_bytes": 145,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "spec_extra_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005348330014385283,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "spec_extra_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005340829957276583,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "spec_extra_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 145,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "439662062674271da71f42300efa3675bce929cc1eb01890cbac32a3692ecd6e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005506669986061752,
      "returned_context_bytes": 145,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "spec_extra_value",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "system_exit",
        "exit_code": 2,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": false,
        "output_policy_correct": false,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 1
        },
        "captured_stdout_bytes": 0,
        "captured_stderr_bytes": 145,
        "captured_stdout_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "captured_stderr_sha256": "439662062674271da71f42300efa3675bce929cc1eb01890cbac32a3692ecd6e",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": false,
      "whole_operation_seconds": 0.0005632499960483983,
      "returned_context_bytes": 145,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "spec_extra_value",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec",
        "SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "arguments",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "arguments",
        "code": "invalid_command_arguments",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 141,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0df08a0c70f4c83b93cb865cd78a86713c0f7c9e45a2308a709a68352c1ff9e7",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005361250077839941,
      "returned_context_bytes": 141,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "root_help",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 709,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "982952f0e5840dc0d6fa0b6c358737891de609ff013d4ff3ecce4d604833b9d1",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007801669999025762,
      "returned_context_bytes": 709,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "root_help",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 709,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "982952f0e5840dc0d6fa0b6c358737891de609ff013d4ff3ecce4d604833b9d1",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006681250088149682,
      "returned_context_bytes": 709,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "root_help",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 709,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "982952f0e5840dc0d6fa0b6c358737891de609ff013d4ff3ecce4d604833b9d1",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000648125002044253,
      "returned_context_bytes": 709,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "root_help",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 709,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "982952f0e5840dc0d6fa0b6c358737891de609ff013d4ff3ecce4d604833b9d1",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006599170010304078,
      "returned_context_bytes": 709,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "root_help",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 709,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "982952f0e5840dc0d6fa0b6c358737891de609ff013d4ff3ecce4d604833b9d1",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000658417004160583,
      "returned_context_bytes": 709,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "root_help",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 709,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "982952f0e5840dc0d6fa0b6c358737891de609ff013d4ff3ecce4d604833b9d1",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006540000031236559,
      "returned_context_bytes": 709,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "run_help",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007364579942077398,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "run_help",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007060409989207983,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "run_help",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007206250011222437,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "run_help",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000710375010385178,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "run_help",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007279169949470088,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "run_help",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007004590006545186,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "help_before_unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help",
        "--unknown"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007625419966643676,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "help_before_unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help",
        "--unknown"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007142079994082451,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "help_before_unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help",
        "--unknown"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008085419976850972,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "help_before_unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help",
        "--unknown"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008927499875426292,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "help_before_unknown_flag",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help",
        "--unknown"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007032089924905449,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "help_before_unknown_flag",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "--help",
        "--unknown"
      ],
      "expected": {
        "kind": "help",
        "exit_code": 0
      },
      "result": {
        "status": "help",
        "exit_code": 0,
        "raised": "SystemExit",
        "terminal_json_exact": false,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 811,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "8aae2f7ffe6f02e5acedca4a503b90e77c61159335f2a03d61cfd2a548224b32",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008602500020060688,
      "returned_context_bytes": 811,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "valid_spec",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec"
      ],
      "expected": {
        "kind": "spec",
        "exit_code": 0
      },
      "result": {
        "status": "spec",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 5308,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "ce49ad96569a21972a8aa46d12183125767564300e39751882c527d043111d0b",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000535915998625569,
      "returned_context_bytes": 5308,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "valid_spec",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec"
      ],
      "expected": {
        "kind": "spec",
        "exit_code": 0
      },
      "result": {
        "status": "spec",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 5805,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "7bd4b9afd4a51ee038e21d0e086f025e5e3fa17f7bbbacca769a77e2b197479b",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005482080014189705,
      "returned_context_bytes": 5805,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "valid_spec",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec"
      ],
      "expected": {
        "kind": "spec",
        "exit_code": 0
      },
      "result": {
        "status": "spec",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 5805,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "7bd4b9afd4a51ee038e21d0e086f025e5e3fa17f7bbbacca769a77e2b197479b",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005581250006798655,
      "returned_context_bytes": 5805,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "valid_spec",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec"
      ],
      "expected": {
        "kind": "spec",
        "exit_code": 0
      },
      "result": {
        "status": "spec",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 5308,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "ce49ad96569a21972a8aa46d12183125767564300e39751882c527d043111d0b",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005407920107245445,
      "returned_context_bytes": 5308,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "valid_spec",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec"
      ],
      "expected": {
        "kind": "spec",
        "exit_code": 0
      },
      "result": {
        "status": "spec",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 5308,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "ce49ad96569a21972a8aa46d12183125767564300e39751882c527d043111d0b",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005610840016743168,
      "returned_context_bytes": 5308,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "valid_spec",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "spec"
      ],
      "expected": {
        "kind": "spec",
        "exit_code": 0
      },
      "result": {
        "status": "spec",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 5805,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "7bd4b9afd4a51ee038e21d0e086f025e5e3fa17f7bbbacca769a77e2b197479b",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007053749868646264,
      "returned_context_bytes": 5805,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "valid_check",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "check",
        "exit_code": 0
      },
      "result": {
        "status": "checked",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": "syntax_and_export_ok",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 187,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "4bf5970f42e97c6f06bdb0a6c6d9420d98cc8d7a08c3438eb2afa5098a3e998d",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006957920122658834,
      "returned_context_bytes": 187,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "valid_check",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "check",
        "exit_code": 0
      },
      "result": {
        "status": "checked",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": "syntax_and_export_ok",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 187,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "4bf5970f42e97c6f06bdb0a6c6d9420d98cc8d7a08c3438eb2afa5098a3e998d",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007247089961310849,
      "returned_context_bytes": 187,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "valid_check",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "check",
        "exit_code": 0
      },
      "result": {
        "status": "checked",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": "syntax_and_export_ok",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 187,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "4bf5970f42e97c6f06bdb0a6c6d9420d98cc8d7a08c3438eb2afa5098a3e998d",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.00070525000046473,
      "returned_context_bytes": 187,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "valid_check",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "check",
        "exit_code": 0
      },
      "result": {
        "status": "checked",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": "syntax_and_export_ok",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 187,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "4bf5970f42e97c6f06bdb0a6c6d9420d98cc8d7a08c3438eb2afa5098a3e998d",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.000766833996749483,
      "returned_context_bytes": 187,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "valid_check",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "check",
        "exit_code": 0
      },
      "result": {
        "status": "checked",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": "syntax_and_export_ok",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 187,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "4bf5970f42e97c6f06bdb0a6c6d9420d98cc8d7a08c3438eb2afa5098a3e998d",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007129169971449301,
      "returned_context_bytes": 187,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "valid_check",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "check-task",
        "$SCRIPT"
      ],
      "expected": {
        "kind": "check",
        "exit_code": 0
      },
      "result": {
        "status": "checked",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": "syntax_and_export_ok",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 187,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "4bf5970f42e97c6f06bdb0a6c6d9420d98cc8d7a08c3438eb2afa5098a3e998d",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007138340006349608,
      "returned_context_bytes": 187,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "valid_init",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW"
      ],
      "expected": {
        "kind": "init",
        "exit_code": 0
      },
      "result": {
        "status": "created",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": true,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 1,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 134,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "7f9e4ff3ea9bd72065a7faf6b662a85dda859b25c9d25d1d0e24afbb4c7ca357",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007965420081745833,
      "returned_context_bytes": 134,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "valid_init",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW"
      ],
      "expected": {
        "kind": "init",
        "exit_code": 0
      },
      "result": {
        "status": "created",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": true,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 1,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 134,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "a7fafa6fd95e15c84d2fc47ef8f43cd975f24e0d17fe477e503ece79a44d2409",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0007902920042397454,
      "returned_context_bytes": 134,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "valid_init",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW"
      ],
      "expected": {
        "kind": "init",
        "exit_code": 0
      },
      "result": {
        "status": "created",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": true,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 1,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 134,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "31611c4bb3b1f729b3fbcd003f11c57bd3a93a3f9a0f9877dc6c21902348497e",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006641660002060235,
      "returned_context_bytes": 134,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "valid_init",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW"
      ],
      "expected": {
        "kind": "init",
        "exit_code": 0
      },
      "result": {
        "status": "created",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": true,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 1,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 134,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "86fa65c88afb0e05baa195aeccc13206f1e06d89ab860da6690c7487df4f0936",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006741669931216165,
      "returned_context_bytes": 134,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "valid_init",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW"
      ],
      "expected": {
        "kind": "init",
        "exit_code": 0
      },
      "result": {
        "status": "created",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": true,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 1,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 134,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "33fe77a5d0fed34c44aaa25e3ff737ca7eec467361f61f0c4f342743fe76c218",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0008266670047305524,
      "returned_context_bytes": 134,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "valid_init",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "init-task",
        "$NEW"
      ],
      "expected": {
        "kind": "init",
        "exit_code": 0
      },
      "result": {
        "status": "created",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": true,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 1,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 134,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "d272fda39e005d25e8e269564c089cd69b5deb3c22813431b23f91da0575dc38",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006570000114152208,
      "returned_context_bytes": 134,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "valid_complete_run",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 318,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "0f42c2a12c60540ed4183fd77c3a3c138707b89e2788e1d60a587c1a088cd4ad",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014748339890502393,
      "returned_context_bytes": 318,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "valid_complete_run",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 317,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "1cc5413e52e28fc10f3dae302c581c1fb38435ddcceed1cf2d5307220b33442c",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013589999871328473,
      "returned_context_bytes": 317,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "valid_complete_run",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 317,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "e32fcef8da379bd28d0cca401ca7df080cb44bed7521d1b33b57fd670b4f0fa3",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013215420040069148,
      "returned_context_bytes": 317,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "valid_complete_run",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 317,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "70ae25ea62a1b8866b01b7f4d60b37f0f1a93fad376a11abcb375b2105101252",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014587909972760826,
      "returned_context_bytes": 317,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "valid_complete_run",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 318,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "235ab025597bf547b4723d43404273d78906e67a2cba663b99c62d08051e7e4c",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016262499993899837,
      "returned_context_bytes": 318,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "valid_complete_run",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 317,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "310f274389378676b673d0f1cab267a826e680e30eb5ca84e7632e16f3489566",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0016147499991348013,
      "returned_context_bytes": 317,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "valid_long_abbreviations",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--go",
        "Observe synthetic completion",
        "--max-s",
        "2",
        "--time",
        "5",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 318,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "72bb53d160f516bdf508f5f063873612ef9609621a653fc570890f74e7dfd3dd",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0015128329978324473,
      "returned_context_bytes": 318,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "valid_long_abbreviations",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--go",
        "Observe synthetic completion",
        "--max-s",
        "2",
        "--time",
        "5",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 318,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "715b7ecca6a044e4d09393a50ec0438cc4ad039250248f5f6337afc374cf24ed",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014885420096106827,
      "returned_context_bytes": 318,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "valid_long_abbreviations",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--go",
        "Observe synthetic completion",
        "--max-s",
        "2",
        "--time",
        "5",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 318,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "6e0b942ed88c4400cfc8b0d912bd3429abc77c0c2867f4f118309c5d026f7a56",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013931669964222237,
      "returned_context_bytes": 318,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "valid_long_abbreviations",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--go",
        "Observe synthetic completion",
        "--max-s",
        "2",
        "--time",
        "5",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 317,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "9f1bed9c7c8e308f35ff49f16f147e1a2dbc23a87f42f80a35d3123ba8dbbe89",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0013792079989798367,
      "returned_context_bytes": 317,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "valid_long_abbreviations",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--go",
        "Observe synthetic completion",
        "--max-s",
        "2",
        "--time",
        "5",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 317,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "72e24a78490c2f4a59c49b60f3820e1528e2719bbf924ef1d63a3ee762d01c3a",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014898340014042333,
      "returned_context_bytes": 317,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "valid_long_abbreviations",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--go",
        "Observe synthetic completion",
        "--max-s",
        "2",
        "--time",
        "5",
        "--archive",
        "$ARCHIVE"
      ],
      "expected": {
        "kind": "run",
        "exit_code": 0
      },
      "result": {
        "status": "done",
        "exit_code": 0,
        "raised": null,
        "terminal_json_exact": true,
        "phase": null,
        "code": null,
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": true,
        "init_output_created": false,
        "journal_created": true,
        "source_unchanged": true,
        "new_file_count": 3,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 318,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "4c3c36b7b7c745f1658a81576b4b6c9e3acc073c96b94a6c99f0e4fc7971a597",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 1,
        "module_spec_attempts": 1,
        "module_exec_attempts": 1,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0014335419982671738,
      "returned_context_bytes": 318,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    },
    {
      "case": "malformed_json_configuration",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs",
        "{SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "configuration",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "configuration",
        "code": "invalid_run_configuration",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 178,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "a4eb34339b6064f022c50991c5035ba00bae2d8929243258cebd21651fb161a3",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005772080039605498,
      "returned_context_bytes": 178,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 1
    },
    {
      "case": "malformed_json_configuration",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs",
        "{SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "configuration",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "configuration",
        "code": "invalid_run_configuration",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 178,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "a4eb34339b6064f022c50991c5035ba00bae2d8929243258cebd21651fb161a3",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005735419981647283,
      "returned_context_bytes": 178,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 1
    },
    {
      "case": "malformed_json_configuration",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs",
        "{SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "configuration",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "configuration",
        "code": "invalid_run_configuration",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 178,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "a4eb34339b6064f022c50991c5035ba00bae2d8929243258cebd21651fb161a3",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005767499969806522,
      "returned_context_bytes": 178,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 2
    },
    {
      "case": "malformed_json_configuration",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs",
        "{SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "configuration",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "configuration",
        "code": "invalid_run_configuration",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 178,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "a4eb34339b6064f022c50991c5035ba00bae2d8929243258cebd21651fb161a3",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005730419943574816,
      "returned_context_bytes": 178,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 2
    },
    {
      "case": "malformed_json_configuration",
      "runtime_version": "0.1.7",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs",
        "{SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "configuration",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "configuration",
        "code": "invalid_run_configuration",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 178,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "a4eb34339b6064f022c50991c5035ba00bae2d8929243258cebd21651fb161a3",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0005994579987600446,
      "returned_context_bytes": 178,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "baseline",
      "repeat": 3
    },
    {
      "case": "malformed_json_configuration",
      "runtime_version": "0.1.8",
      "runtime_loaded_from_requested_source": true,
      "argv_template": [
        "run",
        "$SCRIPT",
        "--goal",
        "Observe synthetic completion",
        "--inputs",
        "{SYNTHETIC_ARGUMENT_VALUE"
      ],
      "expected": {
        "kind": "configuration",
        "exit_code": 2
      },
      "result": {
        "status": "failed",
        "exit_code": 2,
        "raised": null,
        "terminal_json_exact": true,
        "phase": "configuration",
        "code": "invalid_run_configuration",
        "status_metadata_correct": true,
        "output_policy_correct": true,
        "exit_code_correct": true,
        "task_import_marker": false,
        "init_output_created": false,
        "journal_created": false,
        "source_unchanged": true,
        "new_file_count": 0,
        "filesystem_policy_correct": true,
        "source_access_policy_correct": true,
        "streams_restored": true,
        "raw_marker_exposures": {
          "stdout": 0,
          "stderr": 0
        },
        "captured_stdout_bytes": 178,
        "captured_stderr_bytes": 0,
        "captured_stdout_sha256": "a4eb34339b6064f022c50991c5035ba00bae2d8929243258cebd21651fb161a3",
        "captured_stderr_sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855",
        "source_read_calls": 0,
        "module_spec_attempts": 0,
        "module_exec_attempts": 0,
        "native_inference_attempts": 0
      },
      "correct": true,
      "whole_operation_seconds": 0.0006055000121705234,
      "returned_context_bytes": 178,
      "actual_model_usage": {
        "requests": 0,
        "models": {},
        "input_tokens": 0,
        "output_tokens": 0,
        "usage_complete": true
      },
      "arm": "candidate",
      "repeat": 3
    }
  ]
}