examination/topics/qunar-ai-fullstack/ai-tooling/single_choice.json

{
  "topic": "ai-tooling",
  "type": "single_choice",
  "schema_version": "1.0.0",
  "generated": "2026-09-02T00:00:00Z",
  "questions": [
    {
      "id": "sc-001",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "哪个 AI 工具最适合复杂编码任务和项目级重构？",
      "explanation": "Claude Code 专为复杂编码任务设计，具有深度代码理解和多文件编辑能力。",
      "source": null,
      "related": [],
      "options": {
        "A": "Cursor",
        "B": "GitHub Copilot",
        "C": "Claude Code",
        "D": "ChatGPT"
      },
      "answer": "C"
    },
    {
      "id": "sc-002",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "Harness 中状态管理器（State Manager）的核心功能是什么？",
      "explanation": "状态管理器负责任务状态管理、进度追踪和错误恢复。",
      "source": null,
      "related": [],
      "options": {
        "A": "管理用户权限",
        "B": "维护对话历史",
        "C": "追踪任务状态、进度和错误恢复",
        "D": "执行工具调用"
      },
      "answer": "C"
    },
    {
      "id": "sc-003",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "可靠性保障体系的哪一层包含 Sub-agent 交叉验证和 Critic 评分？",
      "explanation": "检测层包含 Sub-agent 交叉验证、Hook 自动检查和 Critic 评分。",
      "source": null,
      "related": [],
      "options": {
        "A": "预防层",
        "B": "检测层",
        "C": "纠正层",
        "D": "监控层"
      },
      "answer": "B"
    },
    {
      "id": "sc-004",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "Definition of Done（完成定义）建立的是什么？",
      "explanation": "Definition of Done 定义任务「完成」的条件，避免主观判断。",
      "source": null,
      "related": [],
      "options": {
        "A": "何时终止 AI 系统",
        "B": "任务开始前必须满足的完成条件",
        "C": "错误处理流程",
        "D": "性能基准"
      },
      "answer": "B"
    },
    {
      "id": "sc-005",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "以下哪项不是 AI 工具选择的原则？",
      "explanation": "工具选择原则为：能力匹配、可控性、可组合性、成本效益。速度优化不在其中。",
      "source": null,
      "related": [],
      "options": {
        "A": "能力匹配",
        "B": "速度优化",
        "C": "可组合性",
        "D": "成本效益"
      },
      "answer": "B"
    },
    {
      "id": "sc-006",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "AI 对代码 Review 的提效倍数是多少？",
      "explanation": "AI 对代码 Review 的效率提升为 6 倍。",
      "source": null,
      "related": [],
      "options": {
        "A": "2x",
        "B": "4x",
        "C": "6x",
        "D": "8x"
      },
      "answer": "C"
    },
    {
      "id": "sc-007",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "权限管理器授权后，下一个执行的组件是什么？",
      "explanation": "授权后执行操作，然后状态管理器更新状态。",
      "source": null,
      "related": [],
      "options": {
        "A": "上下文管理器",
        "B": "工具管理器（执行）",
        "C": "状态管理器",
        "D": "Hook 管理器"
      },
      "answer": "B"
    },
    {
      "id": "sc-008",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "「渐进式信任」方法的目的是什么？",
      "explanation": "渐进式信任：从简单任务开始验证 AI 能力，随着信心增长逐步增加复杂度。",
      "source": null,
      "related": [],
      "options": {
        "A": "立即让 AI 处理所有任务",
        "B": "从简单任务开始验证 AI 能力，逐步增加复杂度",
        "C": "永久限制 AI 只做简单任务",
        "D": "完全移除人工监督"
      },
      "answer": "B"
    },
    {
      "id": "sc-009",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "哪个 AI 工具以广泛的语言支持和低延迟的代码补全著称？",
      "explanation": "GitHub Copilot 以广泛的语言支持和低延迟著称。",
      "source": null,
      "related": [],
      "options": {
        "A": "Claude Code",
        "B": "Cursor",
        "C": "GitHub Copilot",
        "D": "ChatGPT"
      },
      "answer": "C"
    },
    {
      "id": "sc-010",
      "type": "single_choice",
      "difficulty": 3,
      "tags": [
        "ai",
        "tools",
        "harness"
      ],
      "question": "哪种方法论涉及将任务拆分为小的、可验证的子任务？",
      "explanation": "Task Decomposition（任务拆解）将复杂任务拆分为小的、可验证的子任务。",
      "source": null,
      "related": [],
      "options": {
        "A": "Context Engineering",
        "B": "Task Decomposition（任务拆解）",
        "C": "Iterative Development",
        "D": "Template Prompts"
      },
      "answer": "B"
    }
  ]
}