STET

task_detail

validation/gpt-5-3-codex/flux-pr-891/task_detail.v1.json

17126 bytes

Back to adjudication
{
  "version": "v1",
  "materialization": {
    "mode": "canonical"
  },
  "identity": {
    "variant_id": "gpt-5.3-codex",
    "requested_model": "gpt-5.3-codex",
    "model_key": "gpt-5-3-codex",
    "run_id": "2026-02-28__13-38-10__gpt-5-3-codex",
    "task_id": "flux-pr-891",
    "trial_name": "flux-pr-891.1-of-1.2026-02-28__13-38-10__gpt-5-3-codex"
  },
  "decision": {
    "status": "pass",
    "primary_signal": "tests",
    "matrix_status": "pass_with_warn",
    "reason_codes": [
      "primary_tests",
      "tests_passed",
      "equivalence_non_equivalent",
      "review_fail"
    ],
    "flags": [
      "equiv_warn",
      "review_warn"
    ],
    "publish": {
      "include_in_leaderboard": true,
      "weak_signal_risk": false
    },
    "rescue": {
      "candidate": false,
      "eligible": false,
      "decision": "not_candidate",
      "required_artifacts_ok": true
    }
  },
  "tests": {
    "outcome": "pass",
    "strength": "strong",
    "commands": [
      "go test -C v2 ./... -count=1 -timeout=300s"
    ],
    "command_results": [
      {
        "command": "go test -C v2 ./... -count=1 -timeout=300s",
        "gold_outcome": "pass",
        "gold_tb_trial_name": "flux-pr-891.1-of-1.validate-gold-1772291503941928000",
        "gold_failure_mode": "unset",
        "agent_outcome": "pass",
        "agent_tb_trial_name": "flux-pr-891.1-of-1.validate-override-1772291544611770000",
        "agent_failure_mode": "unset",
        "agent_retried": false
      }
    ],
    "partial_score": {
      "score": 1,
      "numerator": 1,
      "denominator": 1,
      "level": "command",
      "provenance": "fallback_command_level",
      "reason": "test_case_detail_unavailable"
    },
    "tb": {
      "trial_name": "flux-pr-891.1-of-1.2026-02-28__13-38-10__gpt-5-3-codex",
      "is_resolved": true,
      "failure_mode": "unset",
      "duration_sec": 77.53102175,
      "time_budget_sec": 1200
    },
    "cache": {
      "cache_hit": false,
      "cache_miss_reason": "missing_pinned_dependencies",
      "environment_group_id": "eg_46b6a6c82a4890b6e6e87c39575fd02c"
    }
  },
  "quality": {
    "equivalence": {
      "status": "used",
      "outcome": "non_equivalent",
      "confidence": 0.85,
      "issue_type": "behavioral",
      "notes": "The patch implements typename validation, quote-unescape fixes, and arena-based error object construction, but it changes core behavior versus intent: validation is done at object level (not per `__typename` selection context), uses a different extension code/message format, and omits source-name-based error reporting. This can miss or mis-handle context-specific typename validation cases that the intended change covers."
    },
    "code_review": {
      "status": "used",
      "signal": "fail",
      "confidence": 0.84,
      "issue_type": "unclear",
      "source_field": "code_review",
      "dimension_scores": {
        "correctness": {
          "score_0_4": 2,
          "confidence_0_1": 0.81,
          "rationale": "The patch adds typename validation, but it does so at object level with a single stored type-context and custom error contract, which can diverge from required per-selection-context validation and expected error code/message behavior."
        },
        "edge_case_handling": {
          "score_0_4": 2,
          "confidence_0_1": 0.76,
          "rationale": "Validation only runs when an object value includes __typename and uses a single stored possible-type set, so mixed/merged selection contexts are not explicitly represented."
        },
        "introduced_bug_risk": {
          "score_0_4": 1,
          "confidence_0_1": 0.88,
          "rationale": "A package-global sync.Map keyed by object pointers is introduced and populated during copies without any cleanup path, creating long-lived metadata accumulation risk."
        },
        "maintainability_idioms": {
          "score_0_4": 1,
          "confidence_0_1": 0.84,
          "rationale": "The patch adds cross-cutting behavior via global side storage and duplicates error-object construction logic inside resolvable instead of centralizing helper behavior, which increases coupling and drift risk."
        }
      },
      "findings": [
        {
          "id": "f1",
          "dimension": "correctness",
          "severity": "major",
          "title": "Typename validation is tied to object-level metadata, not merged selection context",
          "detail": "Possible types are attached once per Object node via SetObjectTypeNameInfo and then validated from that single set. This misses the explicit merge-path handling needed when the same response path is composed from multiple type-conditioned selections, so valid typenames can be rejected or invalid ones missed in merged contexts.",
          "confidence_0_1": 0.82,
          "citations": [
            {
              "path": "v2/pkg/engine/plan/visitor.go",
              "side": "new",
              "start_line": 658,
              "end_line": 670,
              "anchor_kind": "diff",
              "head_sha": "f45cb015e9a28689c83e886b746d21abbd030d37"
            },
            {
              "path": "v2/pkg/engine/resolve/object_typename_info.go",
              "side": "new",
              "start_line": 5,
              "end_line": 9,
              "anchor_kind": "diff",
              "head_sha": "f45cb015e9a28689c83e886b746d21abbd030d37"
            },
            {
              "path": "v2/pkg/engine/resolve/resolvable.go",
              "side": "new",
              "start_line": 987,
              "end_line": 1030,
              "anchor_kind": "diff",
              "head_sha": "f45cb015e9a28689c83e886b746d21abbd030d37"
            }
          ]
        },
        {
          "id": "f2",
          "dimension": "correctness",
          "severity": "major",
          "title": "Introduced error code/message contract differs from intended behavior",
          "detail": "The patch emits a new extension code INVALID_SUBGRAPH_TYPENAME and a different message format. If callers/tests expect the established INVALID_GRAPHQL-style contract for invalid subgraph typename responses, this will fail compatibility despite tests passing locally.",
          "confidence_0_1": 0.79,
          "citations": [
            {
              "path": "v2/pkg/engine/resolve/resolvable.go",
              "side": "new",
              "start_line": 64,
              "end_line": 64,
              "anchor_kind": "diff",
              "head_sha": "f45cb015e9a28689c83e886b746d21abbd030d37"
            },
            {
              "path": "v2/pkg/engine/resolve/resolvable.go",
              "side": "new",
              "start_line": 1015,
              "end_line": 1030,
              "anchor_kind": "diff",
              "head_sha": "f45cb015e9a28689c83e886b746d21abbd030d37"
            }
          ]
        },
        {
          "id": "f3",
          "dimension": "introduced_bug_risk",
          "severity": "major",
          "title": "Global pointer-keyed typename metadata map can leak over time",
          "detail": "objectTypeNameInfos is a package-global sync.Map keyed by *Object. New entries are stored for copied objects, and there is no delete lifecycle. This can retain object graphs indefinitely and increase memory usage under repeated planning/resolution.",
          "confidence_0_1": 0.9,
          "citations": [
            {
              "path": "v2/pkg/engine/resolve/object_typename_info.go",
              "side": "new",
              "start_line": 10,
              "end_line": 33,
              "anchor_kind": "diff",
              "head_sha": "f45cb015e9a28689c83e886b746d21abbd030d37"
            },
            {
              "path": "v2/pkg/engine/resolve/node_object.go",
              "side": "new",
              "start_line": 25,
              "end_line": 28,
              "anchor_kind": "diff",
              "head_sha": "f45cb015e9a28689c83e886b746d21abbd030d37"
            }
          ]
        }
      ],
      "overall": {
        "confidence_overall": 0.84,
        "summary": "The patch implements core pieces and passes provided tests, but it likely does not fully satisfy the intended change due to context-model mismatch for typename validation, contract drift in error code/message, and a high-risk global metadata design."
      },
      "notes": "The patch implements core pieces and passes provided tests, but it likely does not fully satisfy the intended change due to context-model mismatch for typename validation, contract drift in error code/message, and a high-risk global metadata design."
    },
    "behavioral_robustness": {
      "status": "used",
      "probe_status": "used",
      "gold_pass_candidate_pass_count": 1,
      "agreement_rate": 1,
      "coverage_delta_status": "unavailable",
      "coverage_delta_reason": "coverage_adapter_unavailable",
      "mutation_lite_status": "unavailable",
      "mutation_lite_reason": "mutation_runner_unavailable",
      "notes": "derived from command-level outcomes; no accepted generated probes"
    },
    "footprint_risk": {
      "status": "used",
      "reason": "none",
      "version": "v1",
      "level": "low",
      "score": 0.32713156093848017,
      "flag": false,
      "severe_flag": false
    },
    "cost": {
      "status": "tokens_only",
      "token_status": "present",
      "cache_token_status": "present",
      "total_input_tokens": 3577307,
      "total_output_tokens": 17113,
      "total_tokens": 3594420,
      "uncached_input_tokens": 127451,
      "cache_read_input_tokens": 3449856,
      "cached_input_tokens": 3449856,
      "total_cost_usd": 8.113329,
      "token_source": "openai_cached_tokens_usage",
      "pricing_model_key": "gpt-5.3-codex",
      "pricing_mode": "cache_aware"
    }
  },
  "summary": {
    "task": {
      "matrix_status": "pass_with_warn",
      "tests_outcome": "pass",
      "tests_unknown_cause": null,
      "lane_report_lane": null,
      "lane_report_source": null,
      "lane_report_reasons": null,
      "cache_hit": false,
      "cache_miss_reason": "missing_pinned_dependencies",
      "setup_ms_saved": null,
      "pinned_bytes": null,
      "environment_group_id": "eg_46b6a6c82a4890b6e6e87c39575fd02c",
      "equivalence_status": "used",
      "equivalence_outcome": "non_equivalent",
      "code_review_status": "used",
      "code_review_signal": "fail",
      "behavioral_robustness_status": "used",
      "coverage_delta_status": "unavailable",
      "mutation_lite_status": "unavailable",
      "probe_accepted_candidates": null,
      "probe_accepted_commands": null,
      "probe_agreement_rate": 1,
      "probe_gold_pass_candidate_pass_count": 1,
      "probe_gold_pass_candidate_fail_count": null,
      "probe_review_required_count": null,
      "flags": [
        "equiv_warn",
        "review_warn"
      ],
      "tb_is_resolved": true,
      "tb_failure_mode": "unset",
      "tb_total_input_tokens": 3577307,
      "tb_total_output_tokens": 17113,
      "tb_total_tokens": 3594420,
      "tb_uncached_input_tokens": 127451,
      "tb_cache_creation_input_tokens": null,
      "tb_cache_read_input_tokens": 3449856,
      "tb_cached_input_tokens": 3449856,
      "token_status": "present",
      "cache_token_status": "present",
      "token_source": "openai_cached_tokens_usage",
      "task_cost": 8.113329,
      "cost_status": "present",
      "cost_pricing_mode": "cache_aware",
      "pricing_model_key": "gpt-5.3-codex",
      "equiv_rescue_policy": "on",
      "rescue_candidate": false,
      "rescue_eligible": false,
      "rescue_decision": "not_candidate",
      "publish_include_in_leaderboard": true,
      "publish_exclusion_reasons": [],
      "publish_weak_signal_risk": false,
      "tests_only_outcome": 1,
      "rescue_aware_outcome": 1,
      "partial_score": 1,
      "partial_score_numerator": 1,
      "partial_score_denominator": 1,
      "partial_score_level": "command",
      "partial_score_provenance": "fallback_command_level",
      "partial_score_reason": "test_case_detail_unavailable",
      "partial_score_unknown_count": 0,
      "footprint_risk_status": "used",
      "footprint_risk_reason": "none",
      "footprint_risk_level": "low",
      "footprint_risk_score": 0.32713156093848017,
      "footprint_risk_flag": false,
      "footprint_risk_severe_flag": false
    },
    "run": {
      "model": "gpt-5.3-codex",
      "requested_model": "gpt-5.3-codex",
      "run_id": "2026-02-28__13-38-10__gpt-5-3-codex",
      "passRate": 0.9310344827586207,
      "ciLow": 0.8275862068965517,
      "ciHigh": 1,
      "effectiveN": 29,
      "tier": 1,
      "validation_counts": {
        "fail_guardrail": 0,
        "fail_high_conf": 1,
        "fail_infra": 0,
        "fail_likely_equiv": 1,
        "fail_with_diag": 0,
        "missing": 0,
        "pass": 6,
        "pass_with_warn": 21
      }
    },
    "run_metadata": {
      "agent_name": "codex",
      "model_name": "gpt-5.3-codex",
      "start_time": "2026-02-28T13:38:12.733292+00:00",
      "end_time": "2026-02-28T14:43:03.470912+00:00",
      "commit_hash": "99ae882e59281de74856a3e7eeb1b7c986b466af",
      "accuracy": 0.896551724137931,
      "reasoning_effort": "medium"
    }
  },
  "evidence": {
    "task_detail": {
      "path": "validation/gpt-5-3-codex/flux-pr-891/task_detail.v1.json",
      "exists": true
    },
    "trajectory": {
      "path": "validation/gpt-5-3-codex/flux-pr-891/trajectory.v1.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 15290
    },
    "validation": {
      "path": "validation/gpt-5-3-codex/flux-pr-891/validation.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 158818
    },
    "results": {
      "path": "runs/2026-02-28__13-38-10__gpt-5-3-codex/results.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 62248
    },
    "run_metadata": {
      "path": "runs/2026-02-28__13-38-10__gpt-5-3-codex/run_metadata.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 1634
    },
    "agent_patch": {
      "path": "runs/2026-02-28__13-38-10__gpt-5-3-codex/flux-pr-891/flux-pr-891.1-of-1.2026-02-28__13-38-10__gpt-5-3-codex/agent-logs/agent.patch",
      "exists": true,
      "non_empty": true,
      "size_bytes": 13670
    },
    "summary": {
      "path": "reports/summary.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 269968
    },
    "manifest": {
      "path": "manifest.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 677
    },
    "validation_inputs": {
      "required_for_rescue_ok": true,
      "gold_patch": {
        "exists": true,
        "non_empty": true,
        "size_bytes": 10055,
        "sha256": "a0195599ed9d85dd37db5345cc3e3ce6bb0437b570d6cf28e6ceaeadeaf16ba4",
        "required_for_rescue": true
      },
      "repo_tarball": {
        "exists": true,
        "non_empty": true,
        "size_bytes": 4127754,
        "sha256": "1dcc65506817b5e10093dccef87843885126d7d0962a028eb4c705e973d29f7c",
        "required_for_rescue": true
      },
      "agent_patch": {
        "path": "runs/2026-02-28__13-38-10__gpt-5-3-codex/flux-pr-891/flux-pr-891.1-of-1.2026-02-28__13-38-10__gpt-5-3-codex/agent-logs/agent.patch",
        "exists": true,
        "non_empty": true,
        "size_bytes": 13670,
        "sha256": "5a7a7d03ebbcd021c0e6a297e220a1cecca73231ec27cb8fb4615fc3f7d24301",
        "required_for_rescue": true
      }
    }
  },
  "provenance": {
    "identity": [
      {
        "source": "manifest",
        "available": true
      },
      {
        "source": "tb_results",
        "available": true
      }
    ],
    "decision": [
      {
        "source": "validation",
        "available": true
      }
    ],
    "tests": [
      {
        "source": "validation",
        "available": true
      },
      {
        "source": "tb_results",
        "available": true
      }
    ],
    "quality": {
      "equivalence": [
        {
          "source": "validation",
          "available": true
        }
      ],
      "code_review": [
        {
          "source": "validation",
          "available": true
        }
      ],
      "behavioral_robustness": [
        {
          "source": "validation",
          "available": true
        }
      ],
      "footprint_risk": [
        {
          "source": "validation",
          "available": true
        }
      ],
      "cost": [
        {
          "source": "validation",
          "available": true
        },
        {
          "source": "summary",
          "available": true
        }
      ]
    },
    "summary": {
      "task": [
        {
          "source": "summary",
          "available": true
        }
      ],
      "run": [
        {
          "source": "summary",
          "available": true
        }
      ],
      "run_metadata": [
        {
          "source": "run_metadata",
          "available": true
        }
      ]
    },
    "evidence": [
      {
        "source": "filesystem",
        "available": true
      }
    ]
  }
}