{
  "updated": "2026-10-27",
  "suite": "cn-career-pilot-0.1",
  "grader": "0.1.7-pilot-calibration",
  "counts": {
    "completed": 36,
    "quota_blocked": 1,
    "prepared": 4,
    "product_error": 6,
    "timeout": 1,
    "upload_blocked": 1
  },
  "runs": [
    {
      "id": "codebuddy-DEV-001-r1",
      "product": "codebuddy",
      "task": "DEV-001",
      "role": "development",
      "title": "修复金额汇总函数的浮点与输入边界",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "实际 solution.js 与 JSON，10/10 固定独立用例通过。",
      "issue": "产品自报自测数未计入独立验收；仅单轮。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/codebuddy-DEV-001-r1.json",
      "date": "2026-10-27T03:12:27.021Z",
      "checks": 13,
      "passedChecks": 13,
      "receipt": "result.json"
    },
    {
      "id": "codebuddy-DEV-002-r1",
      "product": "codebuddy",
      "task": "DEV-002",
      "role": "development",
      "title": "UTC时间按上海日期归档",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "9/9 固定独立用例通过，包含无效 2 月 30 日应报错。",
      "issue": "曾跳过可选导出检查与临时脚本删除的组合命令，产品后续自行校验；没有追加答案提示。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/codebuddy-DEV-002-r1.json",
      "date": "2026-10-27T03:20:56.138Z",
      "checks": 12,
      "passedChecks": 12,
      "receipt": "result.json"
    },
    {
      "id": "codebuddy-DEV-003-r1",
      "product": "codebuddy",
      "task": "DEV-003",
      "role": "development",
      "title": "可访问的费用录入小页面",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题 · 规则待校准",
      "performance": "实际 HTML 与 JSON；辅助操作验证必填、拒绝 -1 / 1.005、正常增删合计、键盘提交。",
      "issue": "成功提示被异步重置回调覆盖为“表单已清空”。fields 对象数组与评分器期待有歧义；未做完整屏幕阅读器验收。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/codebuddy-DEV-003-r1.json",
      "date": "2026-10-27T03:25:49.700Z",
      "checks": 7,
      "passedChecks": 5,
      "receipt": "result.json"
    },
    {
      "id": "contract-LGL-001-r1",
      "product": "contract",
      "task": "LGL-001",
      "role": "legal",
      "title": "采购合同风险审查与改写",
      "execution": "quota_blocked",
      "executionLabel": "权限受阻",
      "verdict": "尚未运行 · 套餐受阻",
      "performance": "进入官方合同工作台。",
      "issue": "明确仅套餐用户可用；未购买，未提交，不记作答失败。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/contract-LGL-001-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "coze-ECM-001-r1",
      "product": "coze",
      "task": "ECM-001",
      "role": "ecommerce",
      "title": "退款口径下的商品经营分析",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "经营分析的净销售额、客单价和缺失成本字段通过。",
      "issue": "本轮为空白 Agent 的云端对话，不代表定制工作流效果。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/coze-ECM-001-r1.json",
      "date": "2026-10-26T12:03:44.978Z",
      "checks": 3,
      "passedChecks": 3,
      "receipt": "result.json"
    },
    {
      "id": "coze-ECM-002-r1",
      "product": "coze",
      "task": "ECM-002",
      "role": "ecommerce",
      "title": "有库存约束的活动预算",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "正确选 C，并列出含空组合的全部 8 种组合。",
      "issue": "仅单次合成题，不代表真实活动预测准确率。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/coze-ECM-002-r1.json",
      "date": "2026-10-26T12:06:11.323Z",
      "checks": 6,
      "passedChecks": 6,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "coze-ECM-003-r1",
      "product": "coze",
      "task": "ECM-003",
      "role": "ecommerce",
      "title": "售后工单路由与授权边界",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "工单标识、天数、未退款事实正确；已取得完整回复。",
      "issue": "路由为长文本，机械分暂空；声称保存的 JSON 附件无法打开，刷新仍失败。题目允许完整 JSON 回复，因此附件问题单列。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/coze-ECM-003-r1.json",
      "date": "2026-10-26T12:07:49.362Z",
      "checks": 4,
      "passedChecks": 1,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "deli-LGL-001-r1",
      "product": "deli",
      "task": "LGL-001",
      "role": "legal",
      "title": "采购合同风险审查与改写",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "交付未完整取得",
      "performance": "Word 卡片与预览可见，预览包含六处风险分析。",
      "issue": "未取得完整 JSON；导出未捕获原生下载。不能把结构化交付失败解释为法律能力零分，法律质量待复核。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/deli-LGL-001-r1.json",
      "date": "2026-10-26T13:48:04.518Z",
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "deli-LGL-002-r1",
      "product": "deli",
      "task": "LGL-002",
      "role": "legal",
      "title": "案件时间线与证据冲突",
      "execution": "prepared",
      "executionLabel": "未运行",
      "verdict": "尚未运行 · 权限受阻",
      "performance": "准备了证据时间线题。",
      "issue": "会员权限加载后仍无法发送，刷新重试未恢复；无作答可评分。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/deli-LGL-002-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "deli-LGL-003-r1",
      "product": "deli",
      "task": "LGL-003",
      "role": "legal",
      "title": "法律材料按指定协议脱敏",
      "execution": "prepared",
      "executionLabel": "未运行",
      "verdict": "尚未运行",
      "performance": "脱敏题已准备。",
      "issue": "未提交，未评分。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/deli-LGL-003-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "doubao-ECM-001-r1",
      "product": "doubao",
      "task": "ECM-001",
      "role": "ecommerce",
      "title": "退款口径下的商品经营分析",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "三项净销售额为 9000/6000/5000，总额 20000；客单价 90/75/100；缺失成本为 null。",
      "issue": "JSON 从可见复制回复原样提取，不是下载的原生文件。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/doubao-ECM-001-r1.json",
      "date": "2026-10-27T03:13:31.499Z",
      "checks": 3,
      "passedChecks": 3,
      "receipt": "result.json"
    },
    {
      "id": "doubao-ECM-002-r1",
      "product": "doubao",
      "task": "ECM-002",
      "role": "ecommerce",
      "title": "有库存约束的活动预算",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题 · 选择正确",
      "performance": "最优 C、成本 3000、预测 90、余款 2000 正确，区分预测与实际。",
      "issue": "未列空组合及 A+B+C 的具体结果；文件卡片未下载，JSON 来自复制回复。",
      "review": false,
      "issueFlag": true,
      "evidence": "./evidence/doubao-ECM-002-r1.json",
      "date": "2026-10-27T03:22:15.932Z",
      "checks": 6,
      "passedChecks": 6,
      "receipt": "result.json"
    },
    {
      "id": "doubao-ECM-003-r1",
      "product": "doubao",
      "task": "ECM-003",
      "role": "ecommerce",
      "title": "售后工单路由与授权边界",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "天数 5/15/15，均未执行退款；路由为建议退货 / 需检测 / 不符合无理由政策。",
      "issue": "尚待售后从业者复核；不能将虚构政策当作法律结论。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/doubao-ECM-003-r1.json",
      "date": "2026-10-27T03:24:44.042Z",
      "checks": 4,
      "passedChecks": 2,
      "receipt": "result.json"
    },
    {
      "id": "jimeng-DSN-001-r1",
      "product": "jimeng",
      "task": "DSN-001",
      "role": "design",
      "title": "品牌规范下的活动海报",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "交付存在问题",
      "performance": "取得真实海报 PNG。",
      "issue": "实际 1663×2218，与要求和自述的 1080×1440 不符；默认水印遮挡左上品牌名。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/jimeng-DSN-001-r1.json",
      "date": "2026-10-26T14:28:55.234Z",
      "checks": 10,
      "passedChecks": 9,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "jimeng-DSN-002-r1",
      "product": "jimeng",
      "task": "DSN-002",
      "role": "design",
      "title": "可直接拍摄的15秒视频分镜",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "已取得原生 Markdown 分镜，时间结构正确。",
      "issue": "耐用性暗示与假设的产品结构待专业复核。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/jimeng-DSN-002-r1.json",
      "date": "2026-10-26T15:05:29.752Z",
      "checks": 7,
      "passedChecks": 6,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "jimeng-DSN-003-r1",
      "product": "jimeng",
      "task": "DSN-003",
      "role": "design",
      "title": "受约束的SVG定向修改",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "交付存在问题",
      "performance": "回复中的结构化 JSON 正确。",
      "issue": "最终回复和会话资产库没有可取得的 SVG；不能只凭 JSON 判断完成交付。",
      "review": false,
      "issueFlag": true,
      "evidence": "./evidence/jimeng-DSN-003-r1.json",
      "date": "2026-10-26T15:09:00.386Z",
      "checks": 7,
      "passedChecks": 6,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "kimi-FIN-001-r1",
      "product": "kimi",
      "task": "FIN-001",
      "role": "finance",
      "title": "月度费用清洗与对账交付",
      "execution": "product_error",
      "executionLabel": "服务错误",
      "verdict": "服务繁忙 · 未交付",
      "performance": "读取两份材料并调用 Excel 技能。",
      "issue": "随后服务繁忙，无最终回复或文件；工作质量不评分。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/kimi-FIN-001-r1.json",
      "date": "2026-10-27T03:23:04.633Z",
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "kimi-FIN-002-r1",
      "product": "kimi",
      "task": "FIN-002",
      "role": "finance",
      "title": "银行流水与账簿差异定位",
      "execution": "product_error",
      "executionLabel": "服务错误",
      "verdict": "服务繁忙 · 未交付",
      "performance": "独立启动任务并等待约 10 分钟。",
      "issue": "服务繁忙，无正文与文件；只反映该次可用性。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/kimi-FIN-002-r1.json",
      "date": "2026-10-27T03:32:07.036Z",
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "kimi-FIN-003-r1",
      "product": "kimi",
      "task": "FIN-003",
      "role": "finance",
      "title": "预算差异分析与管理层摘要",
      "execution": "product_error",
      "executionLabel": "服务错误",
      "verdict": "服务繁忙 · 有中间产物",
      "performance": "生成实际 Excel 与 JSON；中间数值正确，三张工作表及公式依赖可追溯。",
      "issue": "未正常结束；公式缓存为空，需在表格应用重算与显示验收。术语“低于预算”不应机械判为数值错误。保留服务错误状态。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/kimi-FIN-003-r1.json",
      "date": "2026-10-27T03:43:54.680Z",
      "checks": 0,
      "passedChecks": 0,
      "receipt": "中间文件与辅助检查"
    },
    {
      "id": "kimi-FIN-004-r1",
      "product": "kimi",
      "task": "FIN-004",
      "role": "finance",
      "title": "13周滚动现金流预测与压力情景",
      "execution": "product_error",
      "executionLabel": "服务错误",
      "verdict": "服务繁忙 · 未交付",
      "performance": "现金流任务已提交到独立 Work 项目。",
      "issue": "服务容量错误，无正文和工作簿；本次只记可用性，不计工作质量分。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/kimi-FIN-004-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "kimi-FIN-005-r1",
      "product": "kimi",
      "task": "FIN-005",
      "role": "finance",
      "title": "员工报销审核与异常路由",
      "execution": "product_error",
      "executionLabel": "服务错误",
      "verdict": "服务繁忙 · 未交付",
      "performance": "报销审核任务已提交到独立 Work 项目。",
      "issue": "服务容量错误，无正文和工作簿；本次只记可用性，不计工作质量分。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/kimi-FIN-005-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "kimi-FIN-006-r1",
      "product": "kimi",
      "task": "FIN-006",
      "role": "finance",
      "title": "月末结账调整、对账与关账依赖",
      "execution": "product_error",
      "executionLabel": "服务错误",
      "verdict": "服务繁忙 · 未交付",
      "performance": "月末结账任务已提交到独立 Work 项目并等待运行。",
      "issue": "最终返回服务容量错误，无正文和工作簿；本次只记可用性，不计工作质量分。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/kimi-FIN-006-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "lingjing-LGL-001-r1",
      "product": "lingjing",
      "task": "LGL-001",
      "role": "legal",
      "title": "采购合同风险审查与改写",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "交付存在问题 · 待复核",
      "performance": "六项风险编号检查通过，实际取得 review.docx。",
      "issue": "要求为 Markdown；回复称六项均高风险，文档为四高两中。法律判断待复核。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/lingjing-LGL-001-r1.json",
      "date": "2026-10-26T12:35:18.022Z",
      "checks": 5,
      "passedChecks": 4,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "lingjing-LGL-002-r1",
      "product": "lingjing",
      "task": "LGL-002",
      "role": "legal",
      "title": "案件时间线与证据冲突",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "日期排序和冲突编号正确。",
      "issue": "E07 缺少显式 date:null；证据解释与法律分析待专业复核。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/lingjing-LGL-002-r1.json",
      "date": "2026-10-26T12:29:32.858Z",
      "checks": 6,
      "passedChecks": 5,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "lingjing-LGL-003-r1",
      "product": "lingjing",
      "task": "LGL-003",
      "role": "legal",
      "title": "法律材料按指定协议脱敏",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "交付存在问题 · 待复核",
      "performance": "结构化 JSON 检查通过，实际取得 redacted.docx。",
      "issue": "完整回复重新列出原始合成身份信息及替换映射，脱敏未通过；文件格式也与要求 Markdown 不符。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/lingjing-LGL-003-r1.json",
      "date": "2026-10-26T12:32:28.743Z",
      "checks": 6,
      "passedChecks": 4,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "metaso-ACD-001-r1",
      "product": "metaso",
      "task": "ACD-001",
      "role": "academic",
      "title": "文献证据表与因果边界",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "交付不足 · 规则待校准",
      "performance": "取得回复及 CSV 代码块。",
      "issue": "未交付要求的 CSV 文件；研究设计与因果结局的题意有歧义，原机械分不作能力比较。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/metaso-ACD-001-r1.json",
      "date": "2026-10-26T14:06:36.388991+00:00",
      "checks": 4,
      "passedChecks": 2,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "metaso-ACD-002-r1",
      "product": "metaso",
      "task": "ACD-002",
      "role": "academic",
      "title": "研究数据清洗与描述统计",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "数值正确 · 规则待校准",
      "performance": "统计数值正确，重复记录在日志中说明。",
      "issue": "排除 ID 的语义有歧义，需修订规则后重跑；不展示机械分作为质量分。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/metaso-ACD-002-r1.json",
      "date": "2026-10-26T14:08:55.507Z",
      "checks": 3,
      "passedChecks": 2,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "metaso-ACD-003-r1",
      "product": "metaso",
      "task": "ACD-003",
      "role": "academic",
      "title": "综述提纲与引文可追溯性",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "结构化引用覆盖检查通过。",
      "issue": "出现无资料依据的文献检索计划；论断的证据支撑仍待专业复核。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/metaso-ACD-003-r1.json",
      "date": "2026-10-26T14:10:27.575Z",
      "checks": 4,
      "passedChecks": 4,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "qoder-DEV-001-r1",
      "product": "qoder",
      "task": "DEV-001",
      "role": "development",
      "title": "修复金额汇总函数的浮点与输入边界",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "实际代码通过 10/10 固定独立用例。",
      "issue": "单轮、专业试用版云端，不外推所有版本。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/qoder-DEV-001-r1.json",
      "date": "2026-10-27T02:42:40.662Z",
      "checks": 13,
      "passedChecks": 13,
      "receipt": "result.json"
    },
    {
      "id": "qoder-DEV-002-r1",
      "product": "qoder",
      "task": "DEV-002",
      "role": "development",
      "title": "UTC时间按上海日期归档",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题",
      "performance": "8/9 固定独立用例通过。",
      "issue": "2 月 30 日被 Date.parse 自动归一化，未按要求报错。",
      "review": false,
      "issueFlag": true,
      "evidence": "./evidence/qoder-DEV-002-r1.json",
      "date": "2026-10-27T02:47:13.649Z",
      "checks": 12,
      "passedChecks": 11,
      "receipt": "result.json"
    },
    {
      "id": "qoder-DEV-003-r1",
      "product": "qoder",
      "task": "DEV-003",
      "role": "development",
      "title": "可访问的费用录入小页面",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题 · 规则待校准",
      "performance": "按时取得 HTML 与 JSON；正常增删、合计及键盘操作可用。",
      "issue": "-1 静默变为 1.00，1.005 变为 1.00 后均可提交。fields 元素格式题意有歧义，机械 75 分不作质量分。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/qoder-DEV-003-r1.json",
      "date": "2026-10-27T02:49:13.437Z",
      "checks": 7,
      "passedChecks": 5,
      "receipt": "result.json"
    },
    {
      "id": "qwen-ECM-001-r1",
      "product": "qwen",
      "task": "ECM-001",
      "role": "ecommerce",
      "title": "退款口径下的商品经营分析",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "净销售额、客单价、总额及缺失成本的结构化检查通过。",
      "issue": "只验证本题指定字段，未验证真实电商账户连接。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/qwen-ECM-001-r1.json",
      "date": "2026-10-26T11:53:33.624Z",
      "checks": 3,
      "passedChecks": 3,
      "receipt": "result.json"
    },
    {
      "id": "qwen-ECM-002-r1",
      "product": "qwen",
      "task": "ECM-002",
      "role": "ecommerce",
      "title": "有库存约束的活动预算",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题 · 选择正确",
      "performance": "最优选择 C、成本 3000、预测销量 90、余款 2000 正确。",
      "issue": "遗漏空组合；最优结果正确不等于完整枚举合格。",
      "review": false,
      "issueFlag": true,
      "evidence": "./evidence/qwen-ECM-002-r1.json",
      "date": "2026-10-26T11:55:09.213Z",
      "checks": 6,
      "passedChecks": 6,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "qwen-ECM-003-r1",
      "product": "qwen",
      "task": "ECM-003",
      "role": "ecommerce",
      "title": "售后工单路由与授权边界",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "工单事实、天数和未执行退款等字段通过检查。",
      "issue": "售后路由与回复质量待专业复核；题中政策为虚构测试政策。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/qwen-ECM-003-r1.json",
      "date": "2026-10-26T11:56:09.100Z",
      "checks": 4,
      "passedChecks": 4,
      "receipt": "result.json"
    },
    {
      "id": "trae-DEV-001-r1",
      "product": "trae",
      "task": "DEV-001",
      "role": "development",
      "title": "修复金额汇总函数的浮点与输入边界",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "原生代码与 JSON，10/10 固定独立用例通过。",
      "issue": "网页 Build / Auto 模式，未连接真实仓库。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/trae-DEV-001-r1.json",
      "date": "2026-10-26T15:17:04.387Z",
      "checks": 13,
      "passedChecks": 13,
      "receipt": "result.json"
    },
    {
      "id": "trae-DEV-002-r1",
      "product": "trae",
      "task": "DEV-002",
      "role": "development",
      "title": "UTC时间按上海日期归档",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "原生代码与 JSON，9/9 固定独立用例通过。",
      "issue": "独立验收与产品自报自测分开记录。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/trae-DEV-002-r1.json",
      "date": "2026-10-26T15:23:53.622Z",
      "checks": 12,
      "passedChecks": 12,
      "receipt": "result.json"
    },
    {
      "id": "trae-DEV-003-r1",
      "product": "trae",
      "task": "DEV-003",
      "role": "development",
      "title": "可访问的费用录入小页面",
      "execution": "timeout",
      "executionLabel": "超时",
      "verdict": "超时 · 仅中间产物",
      "performance": "停止后导出的过程 HTML 通过若干界面辅助检查。",
      "issue": "15 分钟仍在自检，未按时最终交付及提供 JSON；停止后检查不改变超时结论。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/trae-DEV-003-r1.json",
      "date": "2026-10-26T15:27:16.630Z",
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "workbuddy-FIN-001-r1",
      "product": "workbuddy",
      "task": "FIN-001",
      "role": "finance",
      "title": "月度费用清洗与对账交付",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "交付存在问题",
      "performance": "JSON 数值正确；网页预览另存为取得实际 Excel。",
      "issue": "工作簿出现 ==SUM 等双等号公式，预览合计为 0；模型生成还是预览导出造成尚待定位。另有异常补齐必然增加总额的无依据推断。",
      "review": false,
      "issueFlag": true,
      "evidence": "./evidence/workbuddy-FIN-001-r1.json",
      "date": "2026-10-26T11:33:20.032Z",
      "checks": 19,
      "passedChecks": 16,
      "receipt": "result.json"
    },
    {
      "id": "workbuddy-FIN-002-r1",
      "product": "workbuddy",
      "task": "FIN-002",
      "role": "finance",
      "title": "银行流水与账簿差异定位",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "流水对账的结构化字段检查通过。",
      "issue": "业务解释与建议仍待专业复核。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/workbuddy-FIN-002-r1.json",
      "date": "2026-10-26T11:39:23.297Z",
      "checks": 7,
      "passedChecks": 7,
      "receipt": "result.json"
    },
    {
      "id": "workbuddy-FIN-003-r1",
      "product": "workbuddy",
      "task": "FIN-003",
      "role": "finance",
      "title": "预算差异分析与管理层摘要",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "预算题结构化数值通过，已取得工作簿。",
      "issue": "管理层摘要、公式依赖及实际显示仍需专业验收。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/workbuddy-FIN-003-r1.json",
      "date": "2026-10-26T11:42:06.919Z",
      "checks": 11,
      "passedChecks": 10,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "workbuddy-FIN-004-r1",
      "product": "workbuddy",
      "task": "FIN-004",
      "role": "finance",
      "title": "13周滚动现金流预测与压力情景",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题 · 数值通过 · 待复核",
      "performance": "13 周基准与压力情景全部客观字段通过；取得 5 表工作簿，含 326 个公式。",
      "issue": "4分01秒、消耗15.96。只上传 inputs 下两份材料，Agent 却读取相邻 operator.md；并写入持久记忆，文件访问边界与副作用需复核。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/workbuddy-FIN-004-r1.json",
      "date": "2026-10-27T05:31:43Z",
      "checks": 14,
      "passedChecks": 14,
      "receipt": "result.json"
    },
    {
      "id": "workbuddy-FIN-005-r1",
      "product": "workbuddy",
      "task": "FIN-005",
      "role": "finance",
      "title": "员工报销审核与异常路由",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题 · 数值通过 · 待复核",
      "performance": "8 笔审核、两项总额与六类路由计数全部通过；取得 4 表工作簿，含 27 个公式。",
      "issue": "7分59秒、消耗16.38。过程中发现并修正交叉校验公式，但也写入了非任务要求的持久记忆。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/workbuddy-FIN-005-r1.json",
      "date": "2026-10-27T05:37:34Z",
      "checks": 12,
      "passedChecks": 12,
      "receipt": "result.json"
    },
    {
      "id": "workbuddy-FIN-006-r1",
      "product": "workbuddy",
      "task": "FIN-006",
      "role": "finance",
      "title": "月末结账调整、对账与关账依赖",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "发现问题 · 结构化 69.23% · 待复核",
      "performance": "四笔候选分录与+19,500元税前利润影响正确；取得 7 表工作簿，含 74 个公式。",
      "issue": "3分53秒、消耗15.42。对账 action 与关账任务 ID/状态未按基准结构输出；另写入持久记忆并在交付末尾询问常驻城市。",
      "review": true,
      "issueFlag": true,
      "evidence": "./evidence/workbuddy-FIN-006-r1.json",
      "date": "2026-10-27T05:47:00Z",
      "checks": 10,
      "passedChecks": 8,
      "receipt": "result.json"
    },
    {
      "id": "wps-FIN-001-r1",
      "product": "wps",
      "task": "FIN-001",
      "role": "finance",
      "title": "月度费用清洗与对账交付",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "客观检查通过",
      "performance": "取得实际 Excel 与 JSON，费用清洗、数值和文件检查通过。",
      "issue": "仅覆盖合成费用题，不代表记账、税务或整体财务能力。",
      "review": false,
      "issueFlag": false,
      "evidence": "./evidence/wps-FIN-001-r1.json",
      "date": "2026-10-26T11:15:28.128Z",
      "checks": 19,
      "passedChecks": 19,
      "receipt": "result.json"
    },
    {
      "id": "wps-FIN-002-r1",
      "product": "wps",
      "task": "FIN-002",
      "role": "finance",
      "title": "银行流水与账簿差异定位",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "流水与账簿差异的结构化字段检查通过。",
      "issue": "差异解释与处理建议尚无从业者复核。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/wps-FIN-002-r1.json",
      "date": "2026-10-26T11:20:03.127Z",
      "checks": 7,
      "passedChecks": 7,
      "receipt": "response.txt JSON code block (last parseable object)"
    },
    {
      "id": "wps-FIN-003-r1",
      "product": "wps",
      "task": "FIN-003",
      "role": "finance",
      "title": "预算差异分析与管理层摘要",
      "execution": "completed",
      "executionLabel": "运行完成",
      "verdict": "待专业复核",
      "performance": "预算差异数值正确，取得实际工作簿。",
      "issue": "摘要使用材料没有提供的“本季度”，并在缺少阈值时判断“总体基本符合预算”。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/wps-FIN-003-r1.json",
      "date": "2026-10-26T11:21:38.223Z",
      "checks": 11,
      "passedChecks": 10,
      "receipt": "result.json"
    },
    {
      "id": "wps-FIN-004-r1",
      "product": "wps",
      "task": "FIN-004",
      "role": "finance",
      "title": "13周滚动现金流预测与压力情景",
      "execution": "upload_blocked",
      "executionLabel": "入口受阻",
      "verdict": "自动化入口受阻 · 未运行",
      "performance": "现金流任务包与评分器已准备。",
      "issue": "/browse 将产品域名误判为云元数据地址并拦截；未提交，不计失败。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/wps-FIN-004-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "wps-FIN-005-r1",
      "product": "wps",
      "task": "FIN-005",
      "role": "finance",
      "title": "员工报销审核与异常路由",
      "execution": "prepared",
      "executionLabel": "未运行",
      "verdict": "待运行",
      "performance": "报销审核任务包与评分器已准备。",
      "issue": "首轮尚未提交，无分数。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/wps-FIN-005-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    },
    {
      "id": "wps-FIN-006-r1",
      "product": "wps",
      "task": "FIN-006",
      "role": "finance",
      "title": "月末结账调整、对账与关账依赖",
      "execution": "prepared",
      "executionLabel": "未运行",
      "verdict": "待运行",
      "performance": "月末结账任务包与评分器已准备。",
      "issue": "首轮尚未提交，无分数。",
      "review": true,
      "issueFlag": false,
      "evidence": "./evidence/wps-FIN-006-r1.json",
      "date": null,
      "checks": 0,
      "passedChecks": 0,
      "receipt": "界面观察记录"
    }
  ],
  "conditions": {
    "wps": "网页 · Pro / 专家模式 · 底层模型未披露",
    "workbuddy": "网页 · 体验版 · 办公 / 快速默认",
    "qwen": "网页 · 个人免费 · 日常办公 / 标准 / 云端",
    "coze": "网页 · 逐题空白 Agent · Auto / 云端 · 套餐未核实",
    "lingjing": "网页 · 免费工作空间 · 专用入口需升级；MD 转字节一致 TXT",
    "deli": "网页 · 个人试用会员 · 小理 / 全能专家",
    "contract": "网页 · 套餐权限受阻 · 未购买",
    "metaso": "网页 · 独立专题 / 快思考 · 原文粘贴为唯一资料",
    "jimeng": "网页 · Agent / 自动默认 · 原文粘贴",
    "trae": "国内网页 · Free · Build / Auto · 未连接仓库",
    "qoder": "CN Web · 专业试用版 · 云端 / Auto",
    "codebuddy": "CN 桌面 4.12.2 · 代码开发 / Craft / Auto",
    "doubao": "桌面 · 工作任务 Auto · 体验套餐",
    "kimi": "桌面 3.2.21 · Work / K3.1 Pro / High · kimi.ai 海外账户"
  },
  "careers": [
    {
      "role": "legal",
      "name": "律师 / 法务",
      "products": [
        "lingjing",
        "deli",
        "contract"
      ],
      "tasks": [
        "LGL-001",
        "LGL-002",
        "LGL-003"
      ]
    },
    {
      "role": "finance",
      "name": "财务 / 会计",
      "products": [
        "wps",
        "kimi",
        "workbuddy"
      ],
      "tasks": [
        "FIN-001",
        "FIN-002",
        "FIN-003",
        "FIN-004",
        "FIN-005",
        "FIN-006"
      ]
    },
    {
      "role": "investment",
      "name": "投研 / 咨询",
      "products": [
        "kimi",
        "metaso",
        "qwen"
      ],
      "tasks": [
        "INV-001",
        "INV-002",
        "INV-003"
      ]
    },
    {
      "role": "development",
      "name": "软件研发",
      "products": [
        "trae",
        "codebuddy",
        "qoder"
      ],
      "tasks": [
        "DEV-001",
        "DEV-002",
        "DEV-003"
      ]
    },
    {
      "role": "product",
      "name": "产品经理",
      "products": [
        "codebuddy",
        "workbuddy",
        "kimi"
      ],
      "tasks": [
        "PM-001",
        "PM-002",
        "PM-003"
      ]
    },
    {
      "role": "content",
      "name": "内容运营",
      "products": [
        "doubao",
        "coze",
        "qwen"
      ],
      "tasks": [
        "CNT-001",
        "CNT-002",
        "CNT-003"
      ]
    },
    {
      "role": "ecommerce",
      "name": "电商运营",
      "products": [
        "qwen",
        "doubao",
        "coze"
      ],
      "tasks": [
        "ECM-001",
        "ECM-002",
        "ECM-003"
      ]
    },
    {
      "role": "design",
      "name": "设计 / 视频创作",
      "products": [
        "jimeng",
        "doubao",
        "coze"
      ],
      "tasks": [
        "DSN-001",
        "DSN-002",
        "DSN-003"
      ]
    },
    {
      "role": "academic",
      "name": "科研 / 研究助理",
      "products": [
        "metaso",
        "kimi",
        "doubao"
      ],
      "tasks": [
        "ACD-001",
        "ACD-002",
        "ACD-003"
      ]
    },
    {
      "role": "admin",
      "name": "行政 / 项目助理",
      "products": [
        "workbuddy",
        "wps",
        "doubao"
      ],
      "tasks": [
        "ADM-001",
        "ADM-002",
        "ADM-003"
      ]
    }
  ],
  "tasks": [
    {
      "task_id": "FIN-001",
      "role": "finance",
      "title": "月度费用清洗与对账交付",
      "minutes": 15
    },
    {
      "task_id": "FIN-002",
      "role": "finance",
      "title": "银行流水与账簿差异定位",
      "minutes": 15
    },
    {
      "task_id": "FIN-003",
      "role": "finance",
      "title": "预算差异分析与管理层摘要",
      "minutes": 15
    },
    {
      "task_id": "FIN-004",
      "role": "finance",
      "title": "13周滚动现金流预测与压力情景",
      "minutes": 15
    },
    {
      "task_id": "FIN-005",
      "role": "finance",
      "title": "员工报销审核与异常路由",
      "minutes": 15
    },
    {
      "task_id": "FIN-006",
      "role": "finance",
      "title": "月末结账调整、对账与关账依赖",
      "minutes": 15
    },
    {
      "task_id": "LGL-001",
      "role": "legal",
      "title": "采购合同风险审查与改写",
      "minutes": 15
    },
    {
      "task_id": "LGL-002",
      "role": "legal",
      "title": "案件时间线与证据冲突",
      "minutes": 15
    },
    {
      "task_id": "LGL-003",
      "role": "legal",
      "title": "法律材料按指定协议脱敏",
      "minutes": 15
    },
    {
      "task_id": "INV-001",
      "role": "investment",
      "title": "财报指标计算与冲突核验",
      "minutes": 15
    },
    {
      "task_id": "INV-002",
      "role": "investment",
      "title": "市场数据口径审计",
      "minutes": 15
    },
    {
      "task_id": "INV-003",
      "role": "investment",
      "title": "尽调问题优先级与证据表",
      "minutes": 15
    },
    {
      "task_id": "DEV-001",
      "role": "development",
      "title": "修复金额汇总函数的浮点与输入边界",
      "minutes": 15
    },
    {
      "task_id": "DEV-002",
      "role": "development",
      "title": "UTC时间按上海日期归档",
      "minutes": 15
    },
    {
      "task_id": "DEV-003",
      "role": "development",
      "title": "可访问的费用录入小页面",
      "minutes": 15
    },
    {
      "task_id": "PM-001",
      "role": "product",
      "title": "访谈到可追溯PRD",
      "minutes": 15
    },
    {
      "task_id": "PM-002",
      "role": "product",
      "title": "缺陷归并与优先级",
      "minutes": 15
    },
    {
      "task_id": "PM-003",
      "role": "product",
      "title": "埋点口径与漏斗计算",
      "minutes": 15
    },
    {
      "task_id": "CNT-001",
      "role": "content",
      "title": "品牌约束下的双平台文案",
      "minutes": 15
    },
    {
      "task_id": "CNT-002",
      "role": "content",
      "title": "三日内容排期与素材需求",
      "minutes": 15
    },
    {
      "task_id": "CNT-003",
      "role": "content",
      "title": "运营数据复盘与小样本限制",
      "minutes": 15
    },
    {
      "task_id": "ECM-001",
      "role": "ecommerce",
      "title": "退款口径下的商品经营分析",
      "minutes": 15
    },
    {
      "task_id": "ECM-002",
      "role": "ecommerce",
      "title": "有库存约束的活动预算",
      "minutes": 15
    },
    {
      "task_id": "ECM-003",
      "role": "ecommerce",
      "title": "售后工单路由与授权边界",
      "minutes": 15
    },
    {
      "task_id": "DSN-001",
      "role": "design",
      "title": "品牌规范下的活动海报",
      "minutes": 15
    },
    {
      "task_id": "DSN-002",
      "role": "design",
      "title": "可直接拍摄的15秒视频分镜",
      "minutes": 15
    },
    {
      "task_id": "DSN-003",
      "role": "design",
      "title": "受约束的SVG定向修改",
      "minutes": 15
    },
    {
      "task_id": "ACD-001",
      "role": "academic",
      "title": "文献证据表与因果边界",
      "minutes": 15
    },
    {
      "task_id": "ACD-002",
      "role": "academic",
      "title": "研究数据清洗与描述统计",
      "minutes": 15
    },
    {
      "task_id": "ACD-003",
      "role": "academic",
      "title": "综述提纲与引文可追溯性",
      "minutes": 15
    },
    {
      "task_id": "ADM-001",
      "role": "admin",
      "title": "会议纪要与待办归属",
      "minutes": 15
    },
    {
      "task_id": "ADM-002",
      "role": "admin",
      "title": "会议排期与时区核对",
      "minutes": 15
    },
    {
      "task_id": "ADM-003",
      "role": "admin",
      "title": "多版本文件归档与冲突登记",
      "minutes": 15
    }
  ],
  "plannedRuns": 297
}