STET

task_detail

validation/gpt-5-1-codex-mini/flux-pr-1351/task_detail.v1.json

14948 bytes

Back to adjudication
{
  "version": "v1",
  "materialization": {
    "mode": "canonical"
  },
  "identity": {
    "variant_id": "gpt-5-1-codex-mini",
    "requested_model": "gpt-5-1-codex-mini",
    "model_key": "gpt-5-1-codex-mini",
    "run_id": "2026-02-21__03-36-16__gpt-5-1-codex-mini",
    "task_id": "flux-pr-1351",
    "trial_name": "flux-pr-1351.1-of-1.2026-02-21__03-36-16__gpt-5-1-codex-mini"
  },
  "decision": {
    "status": "pass",
    "primary_signal": "tests",
    "matrix_status": "pass_with_warn",
    "reason_codes": [
      "primary_tests",
      "tests_passed",
      "decision_conflict",
      "equivalence_non_equivalent",
      "review_unsure"
    ],
    "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-1351.1-of-1.validate-gold-1771694110913177000",
        "gold_failure_mode": "unset",
        "agent_outcome": "pass",
        "agent_tb_trial_name": "flux-pr-1351.1-of-1.validate-override-1771694188476312000",
        "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-1351.1-of-1.2026-02-21__03-36-16__gpt-5-1-codex-mini",
      "is_resolved": true,
      "failure_mode": "unset",
      "duration_sec": 461.08763025,
      "time_budget_sec": 600
    },
    "cache": {
      "cache_hit": false,
      "cache_miss_reason": "missing_pinned_dependencies",
      "environment_group_id": "eg_46b6a6c82a4890b6e6e87c39575fd02c"
    }
  },
  "quality": {
    "equivalence": {
      "status": "used",
      "outcome": "non_equivalent",
      "evaluator_model": "codex-5.3",
      "confidence": 0.88,
      "issue_type": "behavioral",
      "notes": "The patch does add grouped storage and deterministic joining, but it misses core intended behavior for per-subgraph attribution in loader hook responses: `newResponseInfo` call sites now pass `ctx.SubgraphErrors()` (the global joined error), so `ResponseInfo.Err` is still not scoped to the current subgraph. The gold intent is to provide each response with its own subgraph’s error (`subgraphErrors[res.ds.Name]`). Also, the new grouped store appears to be map-backed without clone deep-copy updates, which can cause aliasing across cloned contexts."
    },
    "code_review": {
      "status": "used",
      "signal": "unsure",
      "evaluator_model": "codex-5.3",
      "confidence": 0.88,
      "issue_type": "unclear",
      "source_field": "code_review",
      "dimension_scores": {
        "correctness": {
          "score_0_4": 2,
          "confidence_0_1": 0.92,
          "rationale": "Core grouping was added, but loader hook response construction still appears to pass the globally joined error, which prevents per-subgraph attribution at that integration point."
        },
        "edge_case_handling": {
          "score_0_4": 2,
          "confidence_0_1": 0.8,
          "rationale": "Deterministic ordering and nil filtering are handled, but downstream consumer behavior for per-subgraph vs global error views is not consistently aligned."
        },
        "introduced_bug_risk": {
          "score_0_4": 2,
          "confidence_0_1": 0.86,
          "rationale": "The new store is more complex and broadly wired in, and key call sites now consume aggregated errors, increasing risk of behavioral mismatch despite passing tests."
        },
        "maintainability_idioms": {
          "score_0_4": 3,
          "confidence_0_1": 0.83,
          "rationale": "The refactor is structured and tested, but it introduces an additional error-store abstraction and public grouping API beyond the minimal required change, which increases maintenance surface."
        }
      },
      "findings": [
        {
          "id": "f1",
          "dimension": "correctness",
          "severity": "major",
          "title": "Loader hook still receives globally joined subgraph errors",
          "detail": "`newResponseInfo` call sites were switched to pass `l.ctx.SubgraphErrors()`, which is the aggregate joined error, so `ResponseInfo.Err` can still represent all subgraphs rather than the current result subgraph. This conflicts with the intended per-subgraph attribution behavior.",
          "confidence_0_1": 0.93,
          "citations": [
            {
              "path": "v2/pkg/engine/resolve/loader.go",
              "side": "new",
              "start_line": 252,
              "end_line": 252,
              "anchor_kind": "diff",
              "head_sha": "42024d2d31ac476136dedf2ef63a488b7dd55f63"
            },
            {
              "path": "v2/pkg/engine/resolve/loader.go",
              "side": "new",
              "start_line": 289,
              "end_line": 289,
              "anchor_kind": "diff",
              "head_sha": "42024d2d31ac476136dedf2ef63a488b7dd55f63"
            },
            {
              "path": "v2/pkg/engine/resolve/context.go",
              "side": "new",
              "start_line": 139,
              "end_line": 140,
              "anchor_kind": "diff",
              "head_sha": "42024d2d31ac476136dedf2ef63a488b7dd55f63"
            }
          ]
        }
      ],
      "overall": {
        "confidence_overall": 0.88,
        "summary": "The patch mostly implements grouped subgraph error storage and deterministic ordering, but a key integration point still appears to propagate aggregated errors to loader hook response info, so it likely does not fully satisfy the intended per-subgraph attribution change."
      },
      "notes": "The patch mostly implements grouped subgraph error storage and deterministic ordering, but a key integration point still appears to propagate aggregated errors to loader hook response info, so it likely does not fully satisfy the intended per-subgraph attribution change."
    },
    "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": "medium",
      "score": 0.46109476297018426,
      "flag": false,
      "severe_flag": false
    },
    "cost": {
      "status": "tokens_only",
      "token_status": "present",
      "cache_token_status": "present",
      "total_input_tokens": 4519980,
      "total_output_tokens": 51933,
      "total_tokens": 4571913,
      "uncached_input_tokens": 496044,
      "cache_read_input_tokens": 4023936,
      "cached_input_tokens": 4023936,
      "total_cost_usd": 1.6592544,
      "token_source": "openai_cached_tokens_usage",
      "pricing_model_key": "gpt-5.1-codex-mini",
      "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": "unsure",
      "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": 4519980,
      "tb_total_output_tokens": 51933,
      "tb_total_tokens": 4571913,
      "tb_uncached_input_tokens": 496044,
      "tb_cache_creation_input_tokens": null,
      "tb_cache_read_input_tokens": 4023936,
      "tb_cached_input_tokens": 4023936,
      "token_status": "present",
      "cache_token_status": "present",
      "token_source": "openai_cached_tokens_usage",
      "task_cost": 1.6592544,
      "cost_status": "present",
      "cost_pricing_mode": "cache_aware",
      "pricing_model_key": "gpt-5.1-codex-mini",
      "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": "medium",
      "footprint_risk_score": 0.46109476297018426,
      "footprint_risk_flag": false,
      "footprint_risk_severe_flag": false
    },
    "run": {
      "model": "gpt-5-1-codex-mini",
      "requested_model": "gpt-5-1-codex-mini",
      "run_id": "2026-02-21__03-36-16__gpt-5-1-codex-mini",
      "passRate": 0.5714285714285714,
      "ciLow": 0.38095238095238093,
      "ciHigh": 0.8095238095238095,
      "effectiveN": 21,
      "tier": 1,
      "validation_counts": {
        "fail_guardrail": 13,
        "fail_high_conf": 1,
        "fail_infra": 2,
        "fail_likely_equiv": 0,
        "fail_no_patch": 1,
        "fail_with_diag": 0,
        "missing": 0,
        "pass": 2,
        "pass_with_warn": 11
      }
    },
    "run_metadata": {
      "agent_name": "codex",
      "model_name": "gpt-5.1-codex-mini",
      "start_time": "2026-02-21T03:36:19.601271+00:00",
      "end_time": "2026-02-21T06:29:08.047412+00:00",
      "commit_hash": "f09baebe0424b002663dfd81a4ba49701bb870e3",
      "accuracy": 0.5,
      "reasoning_effort": "medium"
    }
  },
  "evidence": {
    "task_detail": {
      "path": "validation/gpt-5-1-codex-mini/flux-pr-1351/task_detail.v1.json",
      "exists": true
    },
    "trajectory": {
      "path": "validation/gpt-5-1-codex-mini/flux-pr-1351/trajectory.v1.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 15395
    },
    "validation": {
      "path": "validation/gpt-5-1-codex-mini/flux-pr-1351/validation.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 164136
    },
    "results": {
      "path": "runs/2026-02-21__03-36-16__gpt-5-1-codex-mini/results.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 59870
    },
    "run_metadata": {
      "path": "runs/2026-02-21__03-36-16__gpt-5-1-codex-mini/run_metadata.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 1662
    },
    "agent_patch": {
      "path": "runs/2026-02-21__03-36-16__gpt-5-1-codex-mini/flux-pr-1351/flux-pr-1351.1-of-1.2026-02-21__03-36-16__gpt-5-1-codex-mini/agent-logs/agent.patch",
      "exists": true,
      "non_empty": true,
      "size_bytes": 21686
    },
    "summary": {
      "path": "reports/summary.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 96753
    },
    "manifest": {
      "path": "manifest.json",
      "exists": true,
      "non_empty": true,
      "size_bytes": 495
    },
    "validation_inputs": {
      "required_for_rescue_ok": true,
      "gold_patch": {
        "exists": true,
        "non_empty": true,
        "size_bytes": 6203,
        "sha256": "815afccd36fab8f4369b913dc089d46e94ae65c8735d45f3c11f35dc8fd6b631",
        "required_for_rescue": true
      },
      "repo_tarball": {
        "exists": true,
        "non_empty": true,
        "size_bytes": 2164046,
        "sha256": "88c0509275e56d1a37054565c3744cb8706ad963eae4a1ac17b4457ab092b519",
        "required_for_rescue": true
      },
      "agent_patch": {
        "exists": true,
        "non_empty": true,
        "size_bytes": 21686,
        "sha256": "c7e7c97ac9756c2ddaa024932c61773183c0948da735948da7cd76eb5b01fa20",
        "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
      }
    ]
  }
}