EVALARC / DEVELOPMENT AUDIT

Test the grader.
Then trust the signal.

Behavioral controls for task-specific agent evaluations. Correct and faulty submissions are evaluated against the same externally enforced contract.

7/7
declared faults detected
1.000
reference correctness
docker
execution backend

Does the grader detect plausible defects?

Negative controlTargetCandidate scoreResultEvidence
claim-without-actionsrouting0.350Detectedclose-resolved-ticket, retry-after-commit, retry-before-commit, route-open-ticket
wrong-ticketscope0.200Detectedclose-resolved-ticket, retry-after-commit, retry-before-commit, route-open-ticket
wrong-queuerouting0.700Detectedclose-resolved-ticket, retry-after-commit, retry-before-commit, route-open-ticket
duplicate-notenotes0.750Detectedclose-resolved-ticket, retry-after-commit, retry-before-commit, route-open-ticket
close-unresolvedclosure0.900Detectedretry-before-commit, route-open-ticket
skip-retrynotes0.812Detectedretry-before-commit
new-key-on-retrynotes0.938Detectedretry-after-commit

Positive control

routing1.004/4
notes1.004/4
closure1.004/4
scope1.004/4
protocol1.004/4

Reference execution evidence

route-open-ticket · seed 17 · passed
{
  "seed": 17,
  "case_id": "route-open-ticket",
  "checks": {
    "routing": true,
    "notes": true,
    "closure": true,
    "scope": true,
    "protocol": true
  },
  "passed": true,
  "status": "passed",
  "error": null,
  "duration_seconds": 0.802482,
  "transcript_sha256": "7b5b4973e934a60c55b95bda6eedf79fcff62764fbc1b81723c5126608caf4bd",
  "initial_state": {
    "T-85a0bcc1": {
      "category": "technical",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Customer contacted support."
      ]
    },
    "U-6a06e9ab": {
      "category": "billing",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "final_state": {
    "T-85a0bcc1": {
      "category": "technical",
      "queue": "queue-0-2",
      "resolved": false,
      "status": "open",
      "notes": [
        "Customer contacted support.",
        "Reviewed request T-85a0bcc1."
      ]
    },
    "U-6a06e9ab": {
      "category": "billing",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "trace": [
    {
      "step": 0,
      "type": "tool_call",
      "action": {
        "tool": "get_ticket",
        "arguments": {
          "ticket_id": "T-85a0bcc1"
        }
      },
      "result": {
        "ok": true,
        "ticket": {
          "category": "technical",
          "queue": "inbox",
          "resolved": false,
          "status": "open",
          "notes": [
            "Customer contacted support."
          ]
        }
      },
      "status": "ok",
      "error_source": null,
      "changes": {}
    },
    {
      "step": 1,
      "type": "tool_call",
      "action": {
        "tool": "assign_queue",
        "arguments": {
          "ticket_id": "T-85a0bcc1",
          "queue": "queue-0-2"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-85a0bcc1": {
          "before": {
            "category": "technical",
            "queue": "inbox",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "technical",
            "queue": "queue-0-2",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          }
        }
      }
    },
    {
      "step": 2,
      "type": "tool_call",
      "action": {
        "tool": "add_note",
        "arguments": {
          "ticket_id": "T-85a0bcc1",
          "body": "Reviewed request T-85a0bcc1.",
          "request_id": "review"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-85a0bcc1": {
          "before": {
            "category": "technical",
            "queue": "queue-0-2",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "technical",
            "queue": "queue-0-2",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-85a0bcc1."
            ]
          }
        }
      }
    },
    {
      "step": 3,
      "type": "finish",
      "action": {
        "finish": true
      }
    }
  ],
  "tool_calls": 3
}
close-resolved-ticket · seed 17 · passed
{
  "seed": 17,
  "case_id": "close-resolved-ticket",
  "checks": {
    "routing": true,
    "notes": true,
    "closure": true,
    "scope": true,
    "protocol": true
  },
  "passed": true,
  "status": "passed",
  "error": null,
  "duration_seconds": 0.704571,
  "transcript_sha256": "a260768c11b589ce3f58658cac6aae8a01d36e388a47fc47c7995f8d5c36708e",
  "initial_state": {
    "T-c41f9dfd": {
      "category": "billing",
      "queue": "inbox",
      "resolved": true,
      "status": "open",
      "notes": [
        "Customer contacted support."
      ]
    },
    "U-b48438b5": {
      "category": "billing",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "final_state": {
    "T-c41f9dfd": {
      "category": "billing",
      "queue": "queue-1-0",
      "resolved": true,
      "status": "closed",
      "notes": [
        "Customer contacted support.",
        "Reviewed request T-c41f9dfd."
      ]
    },
    "U-b48438b5": {
      "category": "billing",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "trace": [
    {
      "step": 0,
      "type": "tool_call",
      "action": {
        "tool": "get_ticket",
        "arguments": {
          "ticket_id": "T-c41f9dfd"
        }
      },
      "result": {
        "ok": true,
        "ticket": {
          "category": "billing",
          "queue": "inbox",
          "resolved": true,
          "status": "open",
          "notes": [
            "Customer contacted support."
          ]
        }
      },
      "status": "ok",
      "error_source": null,
      "changes": {}
    },
    {
      "step": 1,
      "type": "tool_call",
      "action": {
        "tool": "assign_queue",
        "arguments": {
          "ticket_id": "T-c41f9dfd",
          "queue": "queue-1-0"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-c41f9dfd": {
          "before": {
            "category": "billing",
            "queue": "inbox",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "billing",
            "queue": "queue-1-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          }
        }
      }
    },
    {
      "step": 2,
      "type": "tool_call",
      "action": {
        "tool": "add_note",
        "arguments": {
          "ticket_id": "T-c41f9dfd",
          "body": "Reviewed request T-c41f9dfd.",
          "request_id": "review"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-c41f9dfd": {
          "before": {
            "category": "billing",
            "queue": "queue-1-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "billing",
            "queue": "queue-1-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-c41f9dfd."
            ]
          }
        }
      }
    },
    {
      "step": 3,
      "type": "tool_call",
      "action": {
        "tool": "close_ticket",
        "arguments": {
          "ticket_id": "T-c41f9dfd"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-c41f9dfd": {
          "before": {
            "category": "billing",
            "queue": "queue-1-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-c41f9dfd."
            ]
          },
          "after": {
            "category": "billing",
            "queue": "queue-1-0",
            "resolved": true,
            "status": "closed",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-c41f9dfd."
            ]
          }
        }
      }
    },
    {
      "step": 4,
      "type": "finish",
      "action": {
        "finish": true
      }
    }
  ],
  "tool_calls": 4
}
retry-before-commit · seed 17 · passed
{
  "seed": 17,
  "case_id": "retry-before-commit",
  "checks": {
    "routing": true,
    "notes": true,
    "closure": true,
    "scope": true,
    "protocol": true
  },
  "passed": true,
  "status": "passed",
  "error": null,
  "duration_seconds": 0.710904,
  "transcript_sha256": "b1d1d91dc01fc4bfb3243aa3739e953a40aa3920caf63a4b2ab73c87ecabda66",
  "initial_state": {
    "T-3fb81d27": {
      "category": "technical",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Customer contacted support."
      ]
    },
    "U-62590992": {
      "category": "account",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "final_state": {
    "T-3fb81d27": {
      "category": "technical",
      "queue": "queue-2-1",
      "resolved": false,
      "status": "open",
      "notes": [
        "Customer contacted support.",
        "Reviewed request T-3fb81d27."
      ]
    },
    "U-62590992": {
      "category": "account",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "trace": [
    {
      "step": 0,
      "type": "tool_call",
      "action": {
        "tool": "get_ticket",
        "arguments": {
          "ticket_id": "T-3fb81d27"
        }
      },
      "result": {
        "ok": true,
        "ticket": {
          "category": "technical",
          "queue": "inbox",
          "resolved": false,
          "status": "open",
          "notes": [
            "Customer contacted support."
          ]
        }
      },
      "status": "ok",
      "error_source": null,
      "changes": {}
    },
    {
      "step": 1,
      "type": "tool_call",
      "action": {
        "tool": "assign_queue",
        "arguments": {
          "ticket_id": "T-3fb81d27",
          "queue": "queue-2-1"
        }
      },
      "result": {
        "ok": false,
        "error": {
          "code": "temporarily_unavailable",
          "retryable": true
        }
      },
      "status": "tool_error",
      "error_source": "injected",
      "changes": {}
    },
    {
      "step": 2,
      "type": "tool_call",
      "action": {
        "tool": "assign_queue",
        "arguments": {
          "ticket_id": "T-3fb81d27",
          "queue": "queue-2-1"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-3fb81d27": {
          "before": {
            "category": "technical",
            "queue": "inbox",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "technical",
            "queue": "queue-2-1",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          }
        }
      }
    },
    {
      "step": 3,
      "type": "tool_call",
      "action": {
        "tool": "add_note",
        "arguments": {
          "ticket_id": "T-3fb81d27",
          "body": "Reviewed request T-3fb81d27.",
          "request_id": "review"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-3fb81d27": {
          "before": {
            "category": "technical",
            "queue": "queue-2-1",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "technical",
            "queue": "queue-2-1",
            "resolved": false,
            "status": "open",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-3fb81d27."
            ]
          }
        }
      }
    },
    {
      "step": 4,
      "type": "finish",
      "action": {
        "finish": true
      }
    }
  ],
  "tool_calls": 4
}
retry-after-commit · seed 17 · passed
{
  "seed": 17,
  "case_id": "retry-after-commit",
  "checks": {
    "routing": true,
    "notes": true,
    "closure": true,
    "scope": true,
    "protocol": true
  },
  "passed": true,
  "status": "passed",
  "error": null,
  "duration_seconds": 0.736961,
  "transcript_sha256": "c97b016921c8d8fdcdeffd1ebf1f438d5a29be06732c31c86f99fd087bc0b58f",
  "initial_state": {
    "T-cd68615c": {
      "category": "billing",
      "queue": "inbox",
      "resolved": true,
      "status": "open",
      "notes": [
        "Customer contacted support."
      ]
    },
    "U-51436d1f": {
      "category": "account",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "final_state": {
    "T-cd68615c": {
      "category": "billing",
      "queue": "queue-3-0",
      "resolved": true,
      "status": "closed",
      "notes": [
        "Customer contacted support.",
        "Reviewed request T-cd68615c."
      ]
    },
    "U-51436d1f": {
      "category": "account",
      "queue": "inbox",
      "resolved": false,
      "status": "open",
      "notes": [
        "Unrelated request; preserve this ticket."
      ]
    }
  },
  "trace": [
    {
      "step": 0,
      "type": "tool_call",
      "action": {
        "tool": "get_ticket",
        "arguments": {
          "ticket_id": "T-cd68615c"
        }
      },
      "result": {
        "ok": true,
        "ticket": {
          "category": "billing",
          "queue": "inbox",
          "resolved": true,
          "status": "open",
          "notes": [
            "Customer contacted support."
          ]
        }
      },
      "status": "ok",
      "error_source": null,
      "changes": {}
    },
    {
      "step": 1,
      "type": "tool_call",
      "action": {
        "tool": "assign_queue",
        "arguments": {
          "ticket_id": "T-cd68615c",
          "queue": "queue-3-0"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-cd68615c": {
          "before": {
            "category": "billing",
            "queue": "inbox",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "billing",
            "queue": "queue-3-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          }
        }
      }
    },
    {
      "step": 2,
      "type": "tool_call",
      "action": {
        "tool": "add_note",
        "arguments": {
          "ticket_id": "T-cd68615c",
          "body": "Reviewed request T-cd68615c.",
          "request_id": "review"
        }
      },
      "result": {
        "ok": false,
        "error": {
          "code": "temporarily_unavailable",
          "retryable": true
        }
      },
      "status": "tool_error",
      "error_source": "injected",
      "changes": {
        "T-cd68615c": {
          "before": {
            "category": "billing",
            "queue": "queue-3-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support."
            ]
          },
          "after": {
            "category": "billing",
            "queue": "queue-3-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-cd68615c."
            ]
          }
        }
      }
    },
    {
      "step": 3,
      "type": "tool_call",
      "action": {
        "tool": "add_note",
        "arguments": {
          "ticket_id": "T-cd68615c",
          "body": "Reviewed request T-cd68615c.",
          "request_id": "review"
        }
      },
      "result": {
        "ok": true,
        "replayed": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {}
    },
    {
      "step": 4,
      "type": "tool_call",
      "action": {
        "tool": "close_ticket",
        "arguments": {
          "ticket_id": "T-cd68615c"
        }
      },
      "result": {
        "ok": true
      },
      "status": "ok",
      "error_source": null,
      "changes": {
        "T-cd68615c": {
          "before": {
            "category": "billing",
            "queue": "queue-3-0",
            "resolved": true,
            "status": "open",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-cd68615c."
            ]
          },
          "after": {
            "category": "billing",
            "queue": "queue-3-0",
            "resolved": true,
            "status": "closed",
            "notes": [
              "Customer contacted support.",
              "Reviewed request T-cd68615c."
            ]
          }
        }
      }
    },
    {
      "step": 5,
      "type": "finish",
      "action": {
        "finish": true
      }
    }
  ],
  "tool_calls": 5
}

Reproduction record

Task: support-routing v0.1.0
Seeds: [17]
Grader SHA-256: 2dced2a842369299e6e924db626fedc120071d5731cd9b79202dba7395b90255
Cases SHA-256: b7cd8663c69da9e1715d5c8e1d707514863dcc333fefc57acc3315053ebc85df
Image ID: sha256:78387bc3881b8273120a12ebe6c1ab22b018ccc2c9adf565ae1ac9b536e184ea
Created: 2026-09-14T01:29:39.246926+00:00

Complete JSON evidence for all controls