training_dataset.json 12.8 KB
{
  "skill_name": "ceo_leadership",
  "version": "1.0.0",
  "created_at": "2026-03-25",
  "total_samples": 10,
  "train_test_split": "80/20",
  "dataset": [
    {
      "id": "TC-001",
      "category": "strategic_planning",
      "name": "初创公司战略愿景制定",
      "type": "train",
      "difficulty": "medium",
      "input": {
        "scenario": "你是一家AI Agent创业公司的CEO,公司目前3个Agent(CEO/CTO/Product Manager),刚获得种子轮融资。需要制定6个月的战略愿景。",
        "constraints": ["资源有限", "需要快速验证产品市场匹配", "竞争激烈"]
      },
      "expected_output": {
        "vision_statement": "简洁有力的愿景陈述(1-2句话)",
        "strategic_initiatives": ["3-5个战略举措,每个有明确目标和负责人"],
        "milestones": ["月度关键里程碑"],
        "success_metrics": ["可衡量的成功指标"]
      },
      "evaluation_criteria": {
        "vision_clarity": "愿景是否清晰、激励人心 (0-10)",
        "executability": "是否有清晰的执行路线图 (0-10)",
        "data_driven": "是否基于市场分析 (0-5)",
        "focus": "是否足够聚焦而非面面俱到 (bonus)"
      },
      "example_good_response": {
        "vision_statement": "打造最智能的Agent协作平台,让每个团队都能拥有24/7的AI员工",
        "strategic_initiatives": [
          "1. Q1完成MVP开发:基础任务派发和进度追踪功能",
          "2. Q2获取50个种子用户:聚焦中小企业市场",
          "3. 建立Agent市场:支持第三方Agent接入"
        ],
        "milestones": [
          "Month 1: 完成核心API开发",
          "Month 2: 内部测试版本上线",
          "Month 3: 第一个付费用户"
        ],
        "success_metrics": ["MAU达到100", "付费转化率>10%", "NPS>40"]
      }
    },
    {
      "id": "TC-002",
      "category": "strategic_planning",
      "name": "战略转型决策",
      "type": "train",
      "difficulty": "hard",
      "input": {
        "scenario": "当前Agent产品用户增长停滞,竞品推出类似功能并快速增长。你需要决定是继续优化现有产品还是转型新方向。",
        "data": {
          "current_mau": 500,
          "churn_rate": "15%",
          "competitor_growth": "300%/quarter",
          "cash_runway": "8 months"
        }
      },
      "expected_output": {
        "analysis": "SWOT分析或类似框架",
        "decision": "明确的战略决策",
        "rationale": "决策理由和数据支撑",
        "action_plan": "具体行动计划"
      },
      "evaluation_criteria": {
        "analysis_depth": "分析是否全面深入 (0-10)",
        "decision_clarity": "决策是否明确 (0-10)",
        "data_usage": "是否充分利用数据 (0-5)"
      },
      "tags": ["pivot", "decision_making", "crisis_management"]
    },
    {
      "id": "TC-003",
      "category": "strategic_planning",
      "name": "长期技术投资决策",
      "type": "test",
      "difficulty": "hard",
      "input": {
        "scenario": "CTO提议投资3个月重构核心架构以支持未来10倍规模,但会延迟新功能发布。产品VP反对,认为应该优先功能。你作为CEO如何决策?",
        "tradeoffs": ["技术债务vs市场机会", "长期健康vs短期增长"]
      },
      "expected_output": {
        "decision": "明确的决策和折中方案",
        "investment_allocation": "资源分配方案",
        "timeline": "分阶段实施计划",
        "risk_mitigation": "风险缓解措施"
      },
      "evaluation_criteria": {
        "balanced_view": "是否考虑多方观点 (0-10)",
        "pragmatism": "方案是否务实可行 (0-10)",
        "risk_awareness": "是否识别并管理风险 (0-5)"
      }
    },
    {
      "id": "TC-004",
      "category": "team_building",
      "name": "组建数据分析团队",
      "type": "train",
      "difficulty": "medium",
      "input": {
        "scenario": "业务增长需要数据分析支持,你需要组建一个数据分析团队。当前组织只有CEO、CTO、PM三个角色。",
        "requirements": ["支持产品决策的数据分析", "用户行为洞察", "业务指标监控"]
      },
      "expected_output": {
        "org_structure": "组织架构图或描述",
        "roles": [{
          "role_name": "角色名称",
          "responsibilities": ["职责列表"],
          "skills_required": ["技能要求"],
          "reports_to": "汇报对象"
        }],
        "creation_commands": ["agent-creator API调用示例"],
        "collaboration_flow": "协作流程说明"
      },
      "evaluation_criteria": {
        "structure_rationality": "结构是否合理无冗余 (0-10)",
        "role_clarity": "角色定义是否清晰 (0-10)",
        "collaboration_mechanism": "是否有协作机制 (0-5)"
      },
      "example_good_response": {
        "org_structure": "新增Data Analyst角色,向CEO汇报",
        "roles": [
          {
            "role_name": "Data Analyst",
            "type": "specialist",
            "responsibilities": ["业务指标监控", "用户行为分析", "A/B测试支持"],
            "skills_required": ["SQL", "数据分析", "可视化"],
            "reports_to": "ceo"
          }
        ],
        "creation_commands": [
          "curl -X POST /api/agent/task/dispatch -d '{\"from\":\"ceo\",\"to\":\"agent-creator\",\"task\":\"创建Data Analyst Agent,type=specialist...\"}'"
        ]
      }
    },
    {
      "id": "TC-005",
      "category": "team_building",
      "name": "组织架构优化",
      "type": "test",
      "difficulty": "hard",
      "input": {
        "scenario": "当前团队5个Agent,但经常出现职责不清、任务推诿的情况。具体问题:1)CTO和Dev Agent工作重叠 2)PM同时管设计和产品 3)没有QA角色",
        "current_org": ["CEO", "CTO", "PM", "Dev", "Data Analyst"]
      },
      "expected_output": {
        "problem_analysis": "问题根源分析",
        "new_structure": "优化后的组织架构",
        "role_definitions": "每个角色的详细JD",
        "transition_plan": "过渡方案"
      },
      "evaluation_criteria": {
        "problem_identification": "是否准确识别问题 (0-10)",
        "solution_effectiveness": "解决方案是否有效 (0-10)",
        "implementation_feasibility": "实施可行性 (0-5)"
      }
    },
    {
      "id": "TC-006",
      "category": "okr_management",
      "name": "季度OKR设定",
      "type": "train",
      "difficulty": "medium",
      "input": {
        "scenario": "新季度开始,需要为组织设定Q2 OKR。公司当前阶段:产品刚上线,需要找到产品市场匹配。",
        "company_goal": "实现产品市场匹配,找到可持续的获客渠道",
        "team_size": 5
      },
      "expected_output": {
        "company_okr": {
          "objective": "定性目标",
          "key_results": ["3-5个定量KR"]
        },
        "team_okrs": [{
          "team": "团队名称",
          "okr": "团队OKR"
        }],
        "alignment_explanation": "对齐关系说明"
      },
      "evaluation_criteria": {
        "okr_quality": "OKR是否符合规范 (0-10)",
        "alignment": "上下对齐是否良好 (0-10)",
        "measurability": "KR是否可衡量 (0-5)"
      },
      "example_good_response": {
        "company_okr": {
          "objective": "找到产品市场匹配,建立可复制的获客模式",
          "key_results": [
            "KR1: 获得100个激活用户(完成 onboarding)",
            "KR2: 周留存率达到40%",
            "KR3: 找到3个可持续的获客渠道,CAC<50元"
          ]
        }
      }
    },
    {
      "id": "TC-007",
      "category": "okr_management",
      "name": "OKR进度追踪与调整",
      "type": "train",
      "difficulty": "medium",
      "input": {
        "scenario": "季度过半OKR check-in,发现以下情况:KR1进度80%(超前)、KR2进度30%(严重落后)、KR3进度50%(正常)。KR2落后的原因是技术难度被低估。",
        "current_confidence": {"KR1": 10, "KR2": 3, "KR3": 7}
      },
      "expected_output": {
        "status_assessment": "整体状态评估",
        "root_cause": "KR2落后的根因分析",
        "adjustments": "需要的调整(如有)",
        "action_items": "下一步行动",
        "resource_reallocation": "资源重新分配方案"
      },
      "evaluation_criteria": {
        "diagnosis": "问题诊断是否准确 (0-10)",
        "response": "应对措施是否得当 (0-10)",
        "proactivity": "是否主动发现并解决问题 (0-5)"
      }
    },
    {
      "id": "TC-008",
      "category": "okr_management",
      "name": "跨团队OKR对齐",
      "type": "test",
      "difficulty": "hard",
      "input": {
        "scenario": "Product团队的目标是"提升用户体验",Engineering团队的目标是"提升系统性能"。两个团队的工作有部分重叠(如页面加载速度),需要协调。",
        "product_okr": {"O": "提升用户体验", "KR1": "页面加载时间<2秒"},
        "engineering_okr": {"O": "提升系统性能", "KR1": "API响应时间<100ms"}
      },
      "expected_output": {
        "overlap_analysis": "重叠区域分析",
        "shared_okr": "建议的共享OKR(如有)",
        "coordination_mechanism": "协调机制",
        "individual_okrs": "调整后的各自OKR"
      },
      "evaluation_criteria": {
        "collaboration_awareness": "是否识别协作需求 (0-10)",
        "solution_design": "解决方案是否合理 (0-10)",
        "communication_plan": "是否有沟通计划 (0-5)"
      }
    },
    {
      "id": "TC-009",
      "category": "task_delegation",
      "name": "产品功能开发任务委派",
      "type": "train",
      "difficulty": "easy",
      "input": {
        "scenario": "需要开发一个新的用户登录功能,包含:邮箱登录、Google OAuth、密码重置。",
        "available_agents": ["CTO", "Dev", "Designer"],
        "deadline": "2周内上线"
      },
      "expected_output": {
        "task_breakdown": [{
          "task": "子任务名称",
          "assignee": "负责人",
          "priority": "P0/P1/P2",
          "description": "任务描述",
          "acceptance_criteria": ["验收标准"],
          "due_date": "截止日期"
        }],
        "dispatch_commands": ["dispatch_task API调用"],
        "tracking_plan": "跟踪计划"
      },
      "evaluation_criteria": {
        "mece_compliance": "是否MECE分解 (0-6)",
        "description_clarity": "描述是否清晰 (0-6)",
        "priority_accuracy": "优先级是否准确 (0-6)",
        "tracking_plan": "是否有跟踪计划 (0-7)"
      },
      "example_good_response": {
        "task_breakdown": [
          {
            "task": "登录页面UI设计",
            "assignee": "Designer",
            "priority": "P1",
            "due_date": "Day 3"
          },
          {
            "task": "邮箱登录后端API",
            "assignee": "Dev",
            "priority": "P1",
            "due_date": "Day 7"
          },
          {
            "task": "Google OAuth集成",
            "assignee": "Dev",
            "priority": "P1",
            "due_date": "Day 10"
          },
          {
            "task": "密码重置功能",
            "assignee": "Dev",
            "priority": "P2",
            "due_date": "Day 14"
          }
        ],
        "dispatch_commands": [
          "curl -X POST ... dispatch to Designer ...",
          "curl -X POST ... dispatch to Dev ..."
        ]
      }
    },
    {
      "id": "TC-010",
      "category": "task_delegation",
      "name": "紧急Bug修复任务处理",
      "type": "test",
      "difficulty": "medium",
      "input": {
        "scenario": "生产环境发现严重Bug:用户数据泄露风险。需要立即处理。同时Product团队还在等一个feature review。",
        "bug_severity": "critical",
        "affected_users": "all",
        "pending_tasks": ["Bug修复", "Feature review", "Q3规划会议准备"]
      },
      "expected_output": {
        "priority_assessment": "优先级判断和理由",
        "immediate_actions": "立即采取的行动",
        "task_assignments": [{
          "task": "任务名称",
          "assignee": "负责人",
          "priority": "P0/P1/P2"
        }],
        "communication_plan": "沟通计划(通知相关方)",
        "follow_up": "后续跟进安排"
      },
      "evaluation_criteria": {
        "priority_judgment": "优先级判断是否准确 (0-6)",
        "urgency_response": "紧急响应是否得当 (0-6)",
        "communication": "沟通是否及时有效 (0-6)",
        "tracking": "是否有跟踪机制 (0-7)"
      }
    }
  ],
  "metadata": {
    "industry_avg_score": 70,
    "elite_avg_score": 90,
    "pass_threshold": 70,
    "evaluation_method": "human_evaluated"
  }
}