{
 "plan": {
  "trials": 1,
  "angelo_processes": 1,
  "chats": 1,
  "tasks_per_trial": 7,
  "task_ids": [
   "read_only_question",
   "cube_creation",
   "move_01",
   "resize_02",
   "bevel_03",
   "material_04",
   "question_05"
  ],
  "user_messages": 7,
  "local_preview_accepts": 0,
  "billing": "Each user message is one billed provider turn on the selected connection; a turn can contain several model rounds. Preview accepts are local UI actions, not provider turns. Dependent tasks are skipped, not retried, when their object is missing.",
  "input_isolation": "on",
  "after_failure": "reset",
  "refusal_fallback": "disabled",
  "cache_ttl": "product default",
  "approve_delegation": null,
  "grading": {
   "contract": 4,
   "task_set": "long",
   "tasks": 7,
   "tasks_digest": "1fd04624cfa4b784",
   "gate_version": 2
  },
  "trusted_scripts": false,
  "approval": "default",
  "always_allow_scripts": false,
  "returning_user": {
   "profile": false,
   "new_chat_per_task": false,
   "request_log": false
  },
  "opening_wait_s": 0
 },
 "connection": {
  "id": "<id>",
  "name": "Claude",
  "mode": "claude_subscription",
  "reuse_native": false
 },
 "load_average_at_end": [
  3.20361328125,
  2.453125,
  3.845703125
 ],
 "messages_sent": 7,
 "invalid_trials": {},
 "grading": {
  "passed": 7,
  "failed": 0,
  "ungraded": 0,
  "skipped": 0,
  "total": 7,
  "not_run": 0,
  "contract": 4,
  "answers": {
   "read": 2,
   "parsed_leniently": 0
  },
  "task_set": "long",
  "tasks_digest": "1fd04624cfa4b784",
  "gate_version": 2
 },
 "journeys": {
  "cold_first_message": {
   "definition": "total_ms of a completed request whose runtime started cold (first message of a trial)",
   "samples": 1,
   "values_ms": [
    2906.64
   ],
   "p50_ms": 2906.64,
   "p75_ms": 2906.64
  },
  "warm_follow_up": {
   "definition": "total_ms of a completed request on an already running runtime",
   "samples": 6,
   "values_ms": [
    2769.906,
    2575.26,
    3378.948,
    5609.996,
    2416.346,
    2703.151
   ],
   "p50_ms": 2736.528,
   "p75_ms": 3226.688
  },
  "prompt_to_first_verified_change": {
   "definition": "first_verified_result_ms of a completed edit whose host checks passed",
   "samples": 5,
   "values_ms": [
    1759.033,
    1576.72,
    2371.944,
    4603.489,
    2413.575
   ],
   "p50_ms": 2371.944,
   "p75_ms": 2413.575
  },
  "staged_preview_accept": {
   "definition": "harness time from the human accept call to the committed result, host checks passed",
   "samples": 0,
   "values_ms": []
  }
 },
 "tasks": {
  "read_only_question": {
   "runs": 1,
   "completed": 1,
   "host_checks_passed": 1,
   "skipped": 0,
   "passed": 1,
   "failed": 0,
   "ungraded": 0
  },
  "cube_creation": {
   "runs": 1,
   "completed": 1,
   "host_checks_passed": 1,
   "skipped": 0,
   "passed": 1,
   "failed": 0,
   "ungraded": 0
  },
  "move_01": {
   "runs": 1,
   "completed": 1,
   "host_checks_passed": 1,
   "skipped": 0,
   "passed": 1,
   "failed": 0,
   "ungraded": 0
  },
  "resize_02": {
   "runs": 1,
   "completed": 1,
   "host_checks_passed": 1,
   "skipped": 0,
   "passed": 1,
   "failed": 0,
   "ungraded": 0
  },
  "bevel_03": {
   "runs": 1,
   "completed": 1,
   "host_checks_passed": 1,
   "skipped": 0,
   "passed": 1,
   "failed": 0,
   "ungraded": 0
  },
  "material_04": {
   "runs": 1,
   "completed": 1,
   "host_checks_passed": 1,
   "skipped": 0,
   "passed": 1,
   "failed": 0,
   "ungraded": 0
  },
  "question_05": {
   "runs": 1,
   "completed": 1,
   "host_checks_passed": 1,
   "skipped": 0,
   "passed": 1,
   "failed": 0,
   "ungraded": 0
  }
 },
 "trials": [
  {
   "trial": 1,
   "connection_id": "<id>",
   "binary": "<local path>",
   "version": "5.2.2 LTS",
   "tasks": [
    {
     "trial": 1,
     "task": "read_only_question",
     "kind": "read_only",
     "source": "new read-only question",
     "approved": [],
     "declined": [],
     "harness_fixture": [
      "fixture:probe_object"
     ],
     "environment": {
      "frame_before": 1,
      "input_isolated": true,
      "frame_after": 1,
      "playback": false,
      "external_input_suspected": false
     },
     "chat": "<id>",
     "send_to_idle_ms": 2942.562,
     "receipt": {
      "status": "completed",
      "runtime": "claude",
      "runtime_start": "cold",
      "phases": {
       "checkpoint_ms": 7.356,
       "main_thread_block_ms": 0.702,
       "runtime_start_ms": 685.479,
       "first_text_ms": 2184.278,
       "total_ms": 2906.64
      },
      "counts": {
       "python_fallbacks": 0,
       "provider_turn_requests": 1,
       "model_rounds": 1
      },
      "session_id": "<id>",
      "usage": {
       "input_tokens": 36172,
       "cached_input_tokens": 34543,
       "cache_write_input_tokens": 1627,
       "output_tokens": 114,
       "reasoning_output_tokens": 0,
       "api_ms": 2092,
       "cost_usd": 0.022213,
       "turn_ms": 2116,
       "requests": 1,
       "turns": 1
      },
      "route": {
       "mode": "claude_subscription",
       "requested_model": "claude-opus-5-5",
       "effort": "low",
       "cache_ttl": "1h",
       "cache_gap": "new_session",
       "fast": false,
       "auto": false,
       "single_model": true,
       "escalation": 0,
       "served_models": [
        "claude-opus-5-5"
       ]
      },
      "path": {
       "edits": 0
      }
     },
     "attempts": 1,
     "usage_all_attempts": {
      "input_tokens": 36172,
      "cached_input_tokens": 34543,
      "cache_write_input_tokens": 1627,
      "output_tokens": 114,
      "reasoning_output_tokens": 0,
      "api_ms": 2092,
      "cost_usd": 0.022213,
      "turn_ms": 2116,
      "requests": 1,
      "turns": 1
     },
     "activity": {
      "tools": {},
      "capability_lookups": 0,
      "lookup_rounds": 0,
      "plan_rejections": 0
     },
     "controller_status": "Completed",
     "host_check": {
      "scene_unchanged": true,
      "answer": {
       "form": "mesh_names",
       "expected": [
        [
         "Brass Marker",
         "Cube"
        ]
       ],
       "answered": [
        "Brass Marker",
        "Cube"
       ],
       "evidence_required": null,
       "image_captured": false,
       "supported": true,
       "images_before_answer": 0,
       "evidence_timing": "row order",
       "stated": "Cube, Brass Marker",
       "parsed": "exactly",
       "outcome": "right",
       "truth": null
      }
     },
     "evidence": {
      "images": 0,
      "sources": {},
      "reused": 0,
      "current_scene_camera": 0
     },
     "host_check_passed": true,
     "status": "completed",
     "runtime": "claude",
     "runtime_start": "cold",
     "first_text_ms": 2184.278,
     "first_scene_change_ms": null,
     "first_verified_result_ms": null,
     "total_ms": 2906.64,
     "main_thread_block_ms": 0.702,
     "approval_wait_ms": null,
     "checkpoint_ms": 7.356,
     "provider_turn_requests": 1,
     "model_rounds": 1,
     "model_rounds_scope": "Claude typed-plan rounds",
     "python_fallbacks": 0,
     "fallback_categories": {},
     "provider_usage": {
      "input_tokens": 36172,
      "cached_input_tokens": 34543,
      "cache_write_input_tokens": 1627,
      "output_tokens": 114,
      "reasoning_output_tokens": 0,
      "api_ms": 2092,
      "cost_usd": 0.022213,
      "turn_ms": 2116,
      "requests": 1,
      "turns": 1
     },
     "warm_ups": 0,
     "warm_up_usage": null,
     "grade": "passed"
    },
    {
     "trial": 1,
     "task": "cube_creation",
     "kind": "edit",
     "source": "eval_agent_tasks 01_cube_dimensions_counts",
     "approved": [],
     "declined": [],
     "environment": {
      "frame_before": 1,
      "input_isolated": true,
      "frame_after": 1,
      "playback": false,
      "external_input_suspected": false
     },
     "chat": "<id>",
     "send_to_idle_ms": 2774.893,
     "receipt": {
      "status": "completed",
      "runtime": "claude",
      "runtime_start": "warm",
      "phases": {
       "checkpoint_ms": 9.567,
       "runtime_start_ms": 0.017,
       "main_thread_block_ms": 99.826,
       "queued_wait_ms": 13.183,
       "precondition_ms": 101.475,
       "approval_wait_ms": 0,
       "staging_ms": 2.116,
       "first_native_action_ms": 1746.733,
       "verification_ms": 0.503,
       "commit_ms": 2.962,
       "first_scene_change_ms": 1759.031,
       "first_verified_result_ms": 1759.033,
       "total_ms": 2769.906
      },
      "counts": {
       "python_fallbacks": 0,
       "provider_turn_requests": 1,
       "model_rounds": 1,
       "tool_invocations": 1,
       "precondition_checks": 5,
       "tool_executions": 1,
       "native_steps": 1,
       "verification_samples": 1,
       "native_commits": 1,
       "tool_completions": 1
      },
      "session_id": "<id>",
      "usage": {
       "input_tokens": 36910,
       "cached_input_tokens": 36170,
       "cache_write_input_tokens": 738,
       "output_tokens": 149,
       "reasoning_output_tokens": 0,
       "api_ms": 1604,
       "cost_usd": 0.016126,
       "turn_ms": 1617,
       "requests": 1,
       "turns": 1
      },
      "route": {
       "mode": "claude_subscription",
       "requested_model": "claude-opus-5-5",
       "effort": "low",
       "cache_ttl": "1h",
       "cache_gap": "under_5m",
       "fast": false,
       "auto": false,
       "single_model": true,
       "escalation": 0,
       "served_models": [
        "claude-opus-5-5"
       ]
      },
      "path": {
       "edits": 1,
       "steps": [
        [
         "typed",
         "ok",
         "angelo_scene_program",
         [
          "create_primitive"
         ],
         1
        ]
       ]
      }
     },
     "attempts": 1,
     "usage_all_attempts": {
      "input_tokens": 36910,
      "cached_input_tokens": 36170,
      "cache_write_input_tokens": 738,
      "output_tokens": 149,
      "reasoning_output_tokens": 0,
      "api_ms": 1604,
      "cost_usd": 0.016126,
      "turn_ms": 1617,
      "requests": 1,
      "turns": 1
     },
     "activity": {
      "tools": {
       "angelo_scene_program": 1
      },
      "capability_lookups": 0,
      "lookup_rounds": 0,
      "plan_rejections": 0
     },
     "controller_status": "Completed",
     "host_check": {
      "success": true,
      "passed": true,
      "status": "pass",
      "checks": [
       {
        "index": 0,
        "kind": "dimensions",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "dimensions": [
          2,
          2,
          2
         ],
         "space": "WORLD",
         "scope": "BASE_MESH"
        }
       },
       {
        "index": 1,
        "kind": "mesh_counts",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "vertices": 8,
         "edges": 12,
         "faces": 6,
         "scope": "BASE_MESH"
        }
       },
       {
        "index": 2,
        "kind": "mesh_degenerate",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "degenerate_faces": 0,
         "zero_length_edges": 0,
         "scope": "BASE_MESH_LOCAL"
        }
       },
       {
        "index": 3,
        "kind": "mesh_manifold",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "boundary_edges": 0,
         "nonmanifold_edges": 0,
         "nonmanifold_vertices": 0,
         "wire_edges": 0,
         "loose_vertices": 0,
         "missing_edges": 0,
         "invalid_geometry": false,
         "scope": "BASE_MESH"
        }
       }
      ],
      "truncated": false,
      "read_only": true,
      "scope": "closed_base_data_predicates"
     },
     "evidence": {
      "images": 0,
      "sources": {},
      "reused": 0,
      "current_scene_camera": 0
     },
     "host_check_passed": true,
     "status": "completed",
     "runtime": "claude",
     "runtime_start": "warm",
     "first_text_ms": null,
     "first_scene_change_ms": 1759.031,
     "first_verified_result_ms": 1759.033,
     "total_ms": 2769.906,
     "main_thread_block_ms": 99.826,
     "approval_wait_ms": 0,
     "checkpoint_ms": 9.567,
     "provider_turn_requests": 1,
     "model_rounds": 1,
     "model_rounds_scope": "Claude typed-plan rounds",
     "python_fallbacks": 0,
     "fallback_categories": {},
     "provider_usage": {
      "input_tokens": 36910,
      "cached_input_tokens": 36170,
      "cache_write_input_tokens": 738,
      "output_tokens": 149,
      "reasoning_output_tokens": 0,
      "api_ms": 1604,
      "cost_usd": 0.016126,
      "turn_ms": 1617,
      "requests": 1,
      "turns": 1
     },
     "warm_ups": 0,
     "warm_up_usage": null,
     "grade": "passed"
    },
    {
     "trial": 1,
     "task": "move_01",
     "kind": "edit",
     "source": "long session: move",
     "approved": [],
     "declined": [],
     "environment": {
      "frame_before": 1,
      "input_isolated": true,
      "frame_after": 1,
      "playback": false,
      "external_input_suspected": false
     },
     "chat": "<id>",
     "send_to_idle_ms": 2580.876,
     "receipt": {
      "status": "completed",
      "runtime": "claude",
      "runtime_start": "warm",
      "phases": {
       "checkpoint_ms": 4.073,
       "runtime_start_ms": 0.008,
       "main_thread_block_ms": 11.411,
       "queued_wait_ms": 11.801,
       "precondition_ms": 8.272,
       "approval_wait_ms": 0,
       "staging_ms": 3.579,
       "first_native_action_ms": 1559.941,
       "verification_ms": 1.347,
       "commit_ms": 5.134,
       "first_scene_change_ms": 1576.719,
       "first_verified_result_ms": 1576.72,
       "total_ms": 2575.26
      },
      "counts": {
       "python_fallbacks": 0,
       "provider_turn_requests": 1,
       "model_rounds": 1,
       "tool_invocations": 1,
       "precondition_checks": 5,
       "tool_executions": 1,
       "native_steps": 1,
       "verification_samples": 1,
       "native_commits": 1,
       "tool_completions": 1
      },
      "session_id": "<id>",
      "usage": {
       "input_tokens": 37507,
       "cached_input_tokens": 36908,
       "cache_write_input_tokens": 597,
       "output_tokens": 129,
       "reasoning_output_tokens": 0,
       "api_ms": 1528,
       "cost_usd": 0.014746,
       "turn_ms": 1536,
       "requests": 1,
       "turns": 1
      },
      "route": {
       "mode": "claude_subscription",
       "requested_model": "claude-opus-5-5",
       "effort": "low",
       "cache_ttl": "1h",
       "cache_gap": "under_5m",
       "fast": false,
       "auto": false,
       "single_model": true,
       "escalation": 0,
       "served_models": [
        "claude-opus-5-5"
       ]
      },
      "path": {
       "edits": 1,
       "steps": [
        [
         "typed",
         "ok",
         "angelo_scene_program",
         [
          "transform"
         ],
         1
        ]
       ]
      }
     },
     "attempts": 1,
     "usage_all_attempts": {
      "input_tokens": 37507,
      "cached_input_tokens": 36908,
      "cache_write_input_tokens": 597,
      "output_tokens": 129,
      "reasoning_output_tokens": 0,
      "api_ms": 1528,
      "cost_usd": 0.014746,
      "turn_ms": 1536,
      "requests": 1,
      "turns": 1
     },
     "activity": {
      "tools": {
       "angelo_scene_program": 1
      },
      "capability_lookups": 0,
      "lookup_rounds": 0,
      "plan_rejections": 0
     },
     "controller_status": "Completed",
     "host_check": {
      "success": true,
      "passed": true,
      "status": "pass",
      "checks": [
       {
        "index": 0,
        "kind": "transform",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "location": [
          -2,
          -1,
          0
         ],
         "scope": "local_RNA"
        }
       }
      ],
      "truncated": false,
      "read_only": true,
      "scope": "closed_base_data_predicates"
     },
     "evidence": {
      "images": 0,
      "sources": {},
      "reused": 0,
      "current_scene_camera": 0
     },
     "host_check_passed": true,
     "status": "completed",
     "runtime": "claude",
     "runtime_start": "warm",
     "first_text_ms": null,
     "first_scene_change_ms": 1576.719,
     "first_verified_result_ms": 1576.72,
     "total_ms": 2575.26,
     "main_thread_block_ms": 11.411,
     "approval_wait_ms": 0,
     "checkpoint_ms": 4.073,
     "provider_turn_requests": 1,
     "model_rounds": 1,
     "model_rounds_scope": "Claude typed-plan rounds",
     "python_fallbacks": 0,
     "fallback_categories": {},
     "provider_usage": {
      "input_tokens": 37507,
      "cached_input_tokens": 36908,
      "cache_write_input_tokens": 597,
      "output_tokens": 129,
      "reasoning_output_tokens": 0,
      "api_ms": 1528,
      "cost_usd": 0.014746,
      "turn_ms": 1536,
      "requests": 1,
      "turns": 1
     },
     "warm_ups": 0,
     "warm_up_usage": null,
     "grade": "passed"
    },
    {
     "trial": 1,
     "task": "resize_02",
     "kind": "edit",
     "source": "long session: resize",
     "approved": [],
     "declined": [],
     "environment": {
      "frame_before": 1,
      "input_isolated": true,
      "frame_after": 1,
      "playback": false,
      "external_input_suspected": false
     },
     "chat": "<id>",
     "send_to_idle_ms": 3385.752,
     "receipt": {
      "status": "completed",
      "runtime": "claude",
      "runtime_start": "warm",
      "phases": {
       "checkpoint_ms": 8.703,
       "runtime_start_ms": 0.014,
       "main_thread_block_ms": 13.25,
       "queued_wait_ms": 13.773,
       "precondition_ms": 9.69,
       "approval_wait_ms": 0,
       "staging_ms": 3.916,
       "first_native_action_ms": 2351.513,
       "verification_ms": 1.507,
       "commit_ms": 5.593,
       "first_scene_change_ms": 2371.943,
       "first_verified_result_ms": 2371.944,
       "total_ms": 3378.948
      },
      "counts": {
       "python_fallbacks": 0,
       "provider_turn_requests": 1,
       "model_rounds": 1,
       "tool_invocations": 1,
       "precondition_checks": 5,
       "tool_executions": 1,
       "native_steps": 1,
       "verification_samples": 1,
       "native_commits": 1,
       "tool_completions": 1
      },
      "session_id": "<id>",
      "usage": {
       "input_tokens": 37927,
       "cached_input_tokens": 37505,
       "cache_write_input_tokens": 420,
       "output_tokens": 226,
       "reasoning_output_tokens": 51,
       "api_ms": 2306,
       "cost_usd": 0.015389,
       "turn_ms": 2318,
       "requests": 1,
       "turns": 1
      },
      "route": {
       "mode": "claude_subscription",
       "requested_model": "claude-opus-5-5",
       "effort": "low",
       "cache_ttl": "1h",
       "cache_gap": "under_5m",
       "fast": false,
       "auto": false,
       "single_model": true,
       "escalation": 0,
       "served_models": [
        "claude-opus-5-5"
       ]
      },
      "path": {
       "edits": 1,
       "steps": [
        [
         "typed",
         "ok",
         "angelo_scene_program",
         [
          "transform"
         ],
         1
        ]
       ]
      }
     },
     "attempts": 1,
     "usage_all_attempts": {
      "input_tokens": 37927,
      "cached_input_tokens": 37505,
      "cache_write_input_tokens": 420,
      "output_tokens": 226,
      "reasoning_output_tokens": 51,
      "api_ms": 2306,
      "cost_usd": 0.015389,
      "turn_ms": 2318,
      "requests": 1,
      "turns": 1
     },
     "activity": {
      "tools": {
       "angelo_scene_program": 1
      },
      "capability_lookups": 0,
      "lookup_rounds": 0,
      "plan_rejections": 0
     },
     "controller_status": "Completed",
     "host_check": {
      "success": true,
      "passed": true,
      "status": "pass",
      "checks": [
       {
        "index": 0,
        "kind": "dimensions",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "dimensions": [
          1,
          2,
          1
         ],
         "space": "WORLD",
         "scope": "BASE_MESH"
        }
       }
      ],
      "truncated": false,
      "read_only": true,
      "scope": "closed_base_data_predicates"
     },
     "evidence": {
      "images": 0,
      "sources": {},
      "reused": 0,
      "current_scene_camera": 0
     },
     "host_check_passed": true,
     "status": "completed",
     "runtime": "claude",
     "runtime_start": "warm",
     "first_text_ms": null,
     "first_scene_change_ms": 2371.943,
     "first_verified_result_ms": 2371.944,
     "total_ms": 3378.948,
     "main_thread_block_ms": 13.25,
     "approval_wait_ms": 0,
     "checkpoint_ms": 8.703,
     "provider_turn_requests": 1,
     "model_rounds": 1,
     "model_rounds_scope": "Claude typed-plan rounds",
     "python_fallbacks": 0,
     "fallback_categories": {},
     "provider_usage": {
      "input_tokens": 37927,
      "cached_input_tokens": 37505,
      "cache_write_input_tokens": 420,
      "output_tokens": 226,
      "reasoning_output_tokens": 51,
      "api_ms": 2306,
      "cost_usd": 0.015389,
      "turn_ms": 2318,
      "requests": 1,
      "turns": 1
     },
     "warm_ups": 0,
     "warm_up_usage": null,
     "grade": "passed"
    },
    {
     "trial": 1,
     "task": "bevel_03",
     "kind": "edit",
     "source": "long session: bevel",
     "approved": [],
     "declined": [],
     "environment": {
      "frame_before": 1,
      "input_isolated": true,
      "frame_after": 1,
      "playback": false,
      "external_input_suspected": false
     },
     "chat": "<id>",
     "send_to_idle_ms": 5616.569,
     "receipt": {
      "status": "completed",
      "runtime": "claude",
      "runtime_start": "warm",
      "phases": {
       "checkpoint_ms": 4.081,
       "runtime_start_ms": 0.008,
       "main_thread_block_ms": 14.742,
       "first_text_ms": 3575.883,
       "queued_wait_ms": 10.387,
       "precondition_ms": 10.359,
       "approval_wait_ms": 0,
       "staging_ms": 4.483,
       "first_native_action_ms": 4581.715,
       "verification_ms": 1.801,
       "commit_ms": 6.283,
       "first_scene_change_ms": 4603.488,
       "first_verified_result_ms": 4603.489,
       "total_ms": 5609.996
      },
      "counts": {
       "python_fallbacks": 0,
       "provider_turn_requests": 1,
       "model_rounds": 1,
       "tool_invocations": 1,
       "precondition_checks": 5,
       "tool_executions": 1,
       "native_steps": 1,
       "verification_samples": 1,
       "native_commits": 1,
       "tool_completions": 1
      },
      "session_id": "<id>",
      "usage": {
       "input_tokens": 38469,
       "cached_input_tokens": 37925,
       "cache_write_input_tokens": 542,
       "output_tokens": 458,
       "reasoning_output_tokens": 170,
       "api_ms": 4549,
       "cost_usd": 0.021089,
       "turn_ms": 4557,
       "requests": 1,
       "turns": 1
      },
      "route": {
       "mode": "claude_subscription",
       "requested_model": "claude-opus-5-5",
       "effort": "low",
       "cache_ttl": "1h",
       "cache_gap": "under_5m",
       "fast": false,
       "auto": false,
       "single_model": true,
       "escalation": 0,
       "served_models": [
        "claude-opus-5-5"
       ]
      },
      "path": {
       "edits": 1,
       "steps": [
        [
         "typed",
         "ok",
         "angelo_scene_program",
         [
          "modifier_add"
         ],
         1
        ]
       ]
      }
     },
     "attempts": 1,
     "usage_all_attempts": {
      "input_tokens": 38469,
      "cached_input_tokens": 37925,
      "cache_write_input_tokens": 542,
      "output_tokens": 458,
      "reasoning_output_tokens": 170,
      "api_ms": 4549,
      "cost_usd": 0.021089,
      "turn_ms": 4557,
      "requests": 1,
      "turns": 1
     },
     "activity": {
      "tools": {
       "angelo_scene_program": 1
      },
      "capability_lookups": 0,
      "lookup_rounds": 0,
      "plan_rejections": 0
     },
     "controller_status": "Completed",
     "host_check": {
      "success": true,
      "passed": true,
      "status": "pass",
      "checks": [
       {
        "index": 0,
        "kind": "modifier",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "present": true,
         "type": "BEVEL",
         "values": {
          "width": 0.05000000074505806,
          "segments": 2
         }
        }
       }
      ],
      "truncated": false,
      "read_only": true,
      "scope": "closed_base_data_predicates"
     },
     "evidence": {
      "images": 0,
      "sources": {},
      "reused": 0,
      "current_scene_camera": 0
     },
     "host_check_passed": true,
     "status": "completed",
     "runtime": "claude",
     "runtime_start": "warm",
     "first_text_ms": 3575.883,
     "first_scene_change_ms": 4603.488,
     "first_verified_result_ms": 4603.489,
     "total_ms": 5609.996,
     "main_thread_block_ms": 14.742,
     "approval_wait_ms": 0,
     "checkpoint_ms": 4.081,
     "provider_turn_requests": 1,
     "model_rounds": 1,
     "model_rounds_scope": "Claude typed-plan rounds",
     "python_fallbacks": 0,
     "fallback_categories": {},
     "provider_usage": {
      "input_tokens": 38469,
      "cached_input_tokens": 37925,
      "cache_write_input_tokens": 542,
      "output_tokens": 458,
      "reasoning_output_tokens": 170,
      "api_ms": 4549,
      "cost_usd": 0.021089,
      "turn_ms": 4557,
      "requests": 1,
      "turns": 1
     },
     "warm_ups": 0,
     "warm_up_usage": null,
     "grade": "passed"
    },
    {
     "trial": 1,
     "task": "material_04",
     "kind": "edit",
     "source": "long session: material",
     "approved": [],
     "declined": [],
     "environment": {
      "frame_before": 1,
      "input_isolated": true,
      "frame_after": 1,
      "playback": false,
      "external_input_suspected": false
     },
     "chat": "<id>",
     "send_to_idle_ms": 2418.578,
     "receipt": {
      "status": "completed",
      "runtime": "claude",
      "runtime_start": "warm",
      "phases": {
       "checkpoint_ms": 3.986,
       "runtime_start_ms": 0.007,
       "main_thread_block_ms": 19.922,
       "queued_wait_ms": 17.31,
       "precondition_ms": 12.053,
       "approval_wait_ms": 0,
       "staging_ms": 6.632,
       "first_native_action_ms": 2384.555,
       "verification_ms": 2.547,
       "commit_ms": 9.843,
       "first_scene_change_ms": 2413.573,
       "first_verified_result_ms": 2413.575,
       "total_ms": 2416.346
      },
      "counts": {
       "python_fallbacks": 0,
       "provider_turn_requests": 1,
       "model_rounds": 1,
       "tool_invocations": 1,
       "precondition_checks": 5,
       "tool_executions": 1,
       "native_steps": 1,
       "verification_samples": 1,
       "native_commits": 1,
       "tool_completions": 1
      },
      "session_id": "<id>",
      "usage": {
       "input_tokens": 39278,
       "cached_input_tokens": 38467,
       "cache_write_input_tokens": 809,
       "output_tokens": 179,
       "reasoning_output_tokens": 0,
       "api_ms": 2344,
       "cost_usd": 0.017753,
       "turn_ms": 2352,
       "requests": 1,
       "turns": 1
      },
      "route": {
       "mode": "claude_subscription",
       "requested_model": "claude-opus-5-5",
       "effort": "low",
       "cache_ttl": "1h",
       "cache_gap": "under_5m",
       "fast": false,
       "auto": false,
       "single_model": true,
       "escalation": 0,
       "served_models": [
        "claude-opus-5-5"
       ]
      },
      "path": {
       "edits": 1,
       "steps": [
        [
         "typed",
         "ok",
         "angelo_scene_program",
         [
          "material_principled"
         ],
         1
        ]
       ]
      }
     },
     "attempts": 1,
     "usage_all_attempts": {
      "input_tokens": 39278,
      "cached_input_tokens": 38467,
      "cache_write_input_tokens": 809,
      "output_tokens": 179,
      "reasoning_output_tokens": 0,
      "api_ms": 2344,
      "cost_usd": 0.017753,
      "turn_ms": 2352,
      "requests": 1,
      "turns": 1
     },
     "activity": {
      "tools": {
       "angelo_scene_program": 1
      },
      "capability_lookups": 0,
      "lookup_rounds": 0,
      "plan_rejections": 0
     },
     "controller_status": "Completed",
     "host_check": {
      "success": true,
      "passed": true,
      "status": "pass",
      "checks": [
       {
        "index": 0,
        "kind": "material_assignment",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "assigned_faces": 6,
         "total_faces": 6
        }
       },
       {
        "index": 1,
        "kind": "material_values",
        "target": "$eval",
        "status": "pass",
        "passed": true,
        "observed": {
         "base_color": [
          0.20000000298023224,
          0.800000011920929,
          0.30000001192092896,
          1
         ],
         "roughness": 0.20000000298023224,
         "metallic": 0
        }
       }
      ],
      "truncated": false,
      "read_only": true,
      "scope": "closed_base_data_predicates"
     },
     "evidence": {
      "images": 0,
      "sources": {},
      "reused": 0,
      "current_scene_camera": 0
     },
     "host_check_passed": true,
     "status": "completed",
     "runtime": "claude",
     "runtime_start": "warm",
     "first_text_ms": null,
     "first_scene_change_ms": 2413.573,
     "first_verified_result_ms": 2413.575,
     "total_ms": 2416.346,
     "main_thread_block_ms": 19.922,
     "approval_wait_ms": 0,
     "checkpoint_ms": 3.986,
     "provider_turn_requests": 1,
     "model_rounds": 1,
     "model_rounds_scope": "Claude typed-plan rounds",
     "python_fallbacks": 0,
     "fallback_categories": {},
     "provider_usage": {
      "input_tokens": 39278,
      "cached_input_tokens": 38467,
      "cache_write_input_tokens": 809,
      "output_tokens": 179,
      "reasoning_output_tokens": 0,
      "api_ms": 2344,
      "cost_usd": 0.017753,
      "turn_ms": 2352,
      "requests": 1,
      "turns": 1
     },
     "warm_ups": 0,
     "warm_up_usage": null,
     "grade": "passed"
    },
    {
     "trial": 1,
     "task": "question_05",
     "kind": "read_only",
     "source": "long session: read-only question",
     "approved": [],
     "declined": [],
     "harness_fixture": [
      "fixture:object_offset"
     ],
     "environment": {
      "frame_before": 1,
      "input_isolated": true,
      "frame_after": 1,
      "playback": false,
      "external_input_suspected": false
     },
     "chat": "<id>",
     "send_to_idle_ms": 2733.518,
     "receipt": {
      "status": "completed",
      "runtime": "claude",
      "runtime_start": "warm",
      "phases": {
       "checkpoint_ms": 8.16,
       "runtime_start_ms": 0.011,
       "main_thread_block_ms": 2.007,
       "first_text_ms": 1779.369,
       "total_ms": 2703.151
      },
      "counts": {
       "python_fallbacks": 0,
       "provider_turn_requests": 1,
       "model_rounds": 1
      },
      "session_id": "<id>",
      "usage": {
       "input_tokens": 40099,
       "cached_input_tokens": 39276,
       "cache_write_input_tokens": 821,
       "output_tokens": 196,
       "reasoning_output_tokens": 12,
       "api_ms": 2675,
       "cost_usd": 0.018351,
       "turn_ms": 2686,
       "requests": 1,
       "turns": 1
      },
      "route": {
       "mode": "claude_subscription",
       "requested_model": "claude-opus-5-5",
       "effort": "low",
       "cache_ttl": "1h",
       "cache_gap": "under_5m",
       "fast": false,
       "auto": false,
       "single_model": true,
       "escalation": 0,
       "served_models": [
        "claude-opus-5-5"
       ]
      },
      "path": {
       "edits": 0
      }
     },
     "attempts": 1,
     "usage_all_attempts": {
      "input_tokens": 40099,
      "cached_input_tokens": 39276,
      "cache_write_input_tokens": 821,
      "output_tokens": 196,
      "reasoning_output_tokens": 12,
      "api_ms": 2675,
      "cost_usd": 0.018351,
      "turn_ms": 2686,
      "requests": 1,
      "turns": 1
     },
     "activity": {
      "tools": {},
      "capability_lookups": 0,
      "lookup_rounds": 0,
      "plan_rejections": 0
     },
     "controller_status": "Completed",
     "host_check": {
      "scene_unchanged": true,
      "answer": {
       "form": "box",
       "expected": [
        [
         1,
         2,
         1,
         -1.75,
         -1.5,
         0.75
        ]
       ],
       "answered": [
        1,
        2,
        1,
        -1.75,
        -1.5,
        0.75
       ],
       "evidence_required": null,
       "image_captured": false,
       "supported": true,
       "images_before_answer": 0,
       "evidence_timing": "row order",
       "stated": "1 2 1; LOCATION: -1.75 -1.5 0.75",
       "parsed": "exactly",
       "outcome": "right",
       "truth": {
        "object": "Eval Cube"
       }
      }
     },
     "evidence": {
      "images": 0,
      "sources": {},
      "reused": 0,
      "current_scene_camera": 0
     },
     "host_check_passed": true,
     "status": "completed",
     "runtime": "claude",
     "runtime_start": "warm",
     "first_text_ms": 1779.369,
     "first_scene_change_ms": null,
     "first_verified_result_ms": null,
     "total_ms": 2703.151,
     "main_thread_block_ms": 2.007,
     "approval_wait_ms": null,
     "checkpoint_ms": 8.16,
     "provider_turn_requests": 1,
     "model_rounds": 1,
     "model_rounds_scope": "Claude typed-plan rounds",
     "python_fallbacks": 0,
     "fallback_categories": {},
     "provider_usage": {
      "input_tokens": 40099,
      "cached_input_tokens": 39276,
      "cache_write_input_tokens": 821,
      "output_tokens": 196,
      "reasoning_output_tokens": 12,
      "api_ms": 2675,
      "cost_usd": 0.018351,
      "turn_ms": 2686,
      "requests": 1,
      "turns": 1
     },
     "warm_ups": 0,
     "warm_up_usage": null,
     "grade": "passed"
    }
   ],
   "success": true,
   "grading": {
    "contract": 4,
    "task_set": "long",
    "tasks": 7,
    "tasks_digest": "1fd04624cfa4b784",
    "gate_version": 2,
    "totals": {
     "passed": 7,
     "failed": 0,
     "ungraded": 0,
     "skipped": 0,
     "total": 7,
     "not_run": 0
    },
    "answers": {
     "read": 2,
     "parsed_leniently": 0
    }
   },
   "mode": "claude_subscription",
   "cache_ttl": "product default",
   "trusted_scripts": false,
   "refusal_fallback": "disabled",
   "approve_delegation": null,
   "approval": "AUTO_BOUNDED",
   "approval_policy": "AUTO_BOUNDED with test-only typed allowlist for what still asks",
   "always_allow_scripts": false,
   "opening": {
    "prepared": false,
    "runtime_live": false,
    "bases": [],
    "waited_s": 0.098,
    "wait_limit_s": 0
   },
   "exit_code": 0,
   "elapsed_s": 30.403,
   "log": "<campaign>/smoke-opus-gate7/trial-01/run.log"
  }
 ],
 "method": "real controller in disposable factory-startup GUI processes; per-request local receipts; closed host checks; a run is passed only when a check read its result, an unchecked answer or edit is ungraded; p50/p75 over completed host-verified tasks only",
 "caveats": [
  "Wall-clock latency includes provider network and inference; on a machine running other work it is measured under that load.",
  "Approval waits are the harness's own automatic approvals, not a human's.",
  "Claude typed-plan rounds and Codex turns are different units; Codex internal rounds are not observable."
 ]
}
