feat: 初始化 CS 知识应试强化系统

- 纯 HTML/CSS/JS 单页应用,支持填空题和选择题交互
- 5 个主题共 100 道题(GC/JVM、AI 结构化输出、记忆管理、幻觉、工具链)
- 按主题组织、按题型拆分的 JSON 数据结构
- JSON Schema 校验 + 提示词模板 + 题型模板
- 自定义导入功能(粘贴/上传 JSON)
- 项目文档(需求 + 架构)

Co-Authored-By: Claude Code <noreply@anthropic.com>
This commit is contained in:
2026-09-02 18:22:20 +08:00
parent cf589b12f4
commit d3cba0e23c
30 changed files with 4092 additions and 1 deletions
+188
View File
@@ -0,0 +1,188 @@
{
"topic": "ai-tooling",
"type": "fill_blank",
"schema_version": "1.0.0",
"generated": "2026-09-02T00:00:00Z",
"questions": [
{
"id": "fb-001",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "Claude Code 是一个______ Agent,具有深度代码理解、工具调用和多文件编辑能力。",
"explanation": "Claude Code 是 CLI(命令行界面)Agent,专为复杂编码任务设计。",
"source": null,
"related": [],
"answer": [
"CLI"
],
"answer_rule": "any"
},
{
"id": "fb-002",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "Harness 中的______管理器决定哪些操作需要用户确认,哪些可以自动执行。",
"explanation": "权限管理器(Permission Manager)控制哪些操作需要用户授权。",
"source": null,
"related": [],
"answer": [
"权限(Permission)"
],
"answer_rule": "any"
},
{
"id": "fb-003",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "可靠性保障体系的三个层次是:预防、检测和______。",
"explanation": "可靠性体系由预防(Prevention)、检测(Detection)和纠正(Correction)三层组成。",
"source": null,
"related": [],
"answer": [
"纠正"
],
"answer_rule": "any"
},
{
"id": "fb-004",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "Harness 中的______ Hook 在工具执行前检查参数合理性。",
"explanation": "preToolUse Hook 在工具执行前验证参数合理性。",
"source": null,
"related": [],
"answer": [
"preToolUse"
],
"answer_rule": "any"
},
{
"id": "fb-005",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "根据面试笔记,AI 对重复性 CRUD 代码的提效倍数是______。",
"explanation": "AI 对 CRUD 代码的效率提升为 6 倍。",
"source": null,
"related": [],
"answer": [
"6x"
],
"answer_rule": "any"
},
{
"id": "fb-006",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "Harness 中的______管理器维护对话历史、工具结果和文件状态。",
"explanation": "上下文管理器(Context Manager)维护完整的对话上下文。",
"source": null,
"related": [],
"answer": [
"上下文(Context)"
],
"answer_rule": "any"
},
{
"id": "fb-007",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "工具选择原则中,______性意味着工具应具备良好的权限控制和审计能力。",
"explanation": "可控性(Controllability)强调工具应有良好的权限控制和审计能力。",
"source": null,
"related": [],
"answer": [
"可控"
],
"answer_rule": "any"
},
{
"id": "fb-008",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "Harness 中的______ Hook 在代码提交前运行 lint、测试和类型检查。",
"explanation": "preCommit Hook 在代码提交前运行自动化检查。",
"source": null,
"related": [],
"answer": [
"preCommit"
],
"answer_rule": "any"
},
{
"id": "fb-009",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "在 Harness 流程中,LLM 处理输入后,下一个组件是______管理器。",
"explanation": "流程为:LLM → 工具管理器 → 权限管理器 → 执行。",
"source": null,
"related": [],
"answer": [
"工具(Tool)"
],
"answer_rule": "any"
},
{
"id": "fb-010",
"type": "fill_blank",
"difficulty": 2,
"tags": [
"ai",
"tools",
"harness"
],
"question": "______工程的核心是为 AI 系统提供精准的上下文。",
"explanation": "Context Engineering(上下文工程)专注于为 AI 提供精准上下文以提升表现。",
"source": null,
"related": [],
"answer": [
"Context"
],
"answer_rule": "any"
}
]
}
+28
View File
@@ -0,0 +1,28 @@
{
"slug": "ai-tooling",
"name": "AI 工具 / Harness / 综合",
"description": "AI 工具选型、Harness 架构、可靠性保障、提效分析",
"tags": [
"ai",
"tools",
"harness",
"claude-code",
"reliability"
],
"difficulty_range": [
1,
5
],
"schema_version": "1.0.0",
"question_files": [
"fill_blank",
"single_choice"
],
"stats": {
"total": 20,
"by_type": {
"fill_blank": 10,
"single_choice": 10
}
}
}
+218
View File
@@ -0,0 +1,218 @@
{
"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"
}
]
}