Back to Details

hw-interview

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

/hw-interview — 模拟面试助手

/hw-interview — Mock Interview Assistant

根据用户简历，结合本包 
knowledge/
生成一套可演练的华为校招技术面题（不依赖任何线上 API）。
Generate a set of practice technical interview questions for Huawei campus recruitment based on the user's resume and combined with the 
knowledge/
directory of this package (no dependency on any online APIs).

知识库位置（强制）

Knowledge Base Location (Mandatory)

  • 知识库在本 skill 根目录下的 
    knowledge/
    （与本 
    SKILL.md
    同级），随 skill 一起分发/安装。
  • 所有 Read / Grep 路径均相对本 skill 根：如 
    knowledge/experiences/platform/
    、
    knowledge/coding-problems/hot100/index.json
    。
  • 禁止到用户当前工作区根目录、其他项目或仓外路径查找 
    knowledge/
    ；找不到本 skill 内知识库时，说明无法出题依据，不要编造面经或题链。
  • The knowledge base is located in 
    knowledge/
    at the root directory of this skill (same level as this 
    SKILL.md
    ), distributed/installed along with the skill.
  • All Read / Grep paths are relative to the root of this skill: e.g., 
    knowledge/experiences/platform/
    , 
    knowledge/coding-problems/hot100/index.json
    .
  • Prohibited to look for 
    knowledge/
    in the user's current workspace root, other projects, or paths outside the repository; if the knowledge base within this skill cannot be found, it indicates there is no basis for generating questions, do not fabricate interview experiences or question chains.

输入

Input

必需：简历（粘贴正文，或用户附带的简历文件文本）。缺简历时只追问一次。
忽略手机号 / 邮箱 / 身份证等隐私字段，不要写入输出。
可选：目标岗位方向（通用软件开发 | AI | 嵌入式 | 通信/网络 | 测试）。未给出时从简历技能与项目推断。
Required: Resume (paste the text, or the text from the resume file attached by the user). Only follow up once if the resume is missing.
Ignore privacy fields such as phone number / email / ID card, do not include them in the output.
Optional: Target job direction (General Software Development | AI | Embedded | Communication/Network | Testing). If not provided, infer from the skills and projects in the resume.

岗位方向（内部）

Job Directions (Internal)

方向简历信号（示例）
通用软件开发Java / C++ / 前端 / 后端 / Spring / 微服务 / 数据库
AI算法 / 机器学习 / 深度学习 / 大模型 / NLP / CV / PyTorch
嵌入式嵌入式 / 单片机 / FPGA / STM32 / RTOS
通信/网络通信 / 5G / 数通 / 协议 / 射频
测试测试 / 测开 / 自动化测试 / QA
DirectionResume Signals (Examples)
General Software DevelopmentJava / C++ / Frontend / Backend / Spring / Microservices / Database
AIAlgorithms / Machine Learning / Deep Learning / Large Models / NLP / CV / PyTorch
EmbeddedEmbedded / Microcontroller / FPGA / STM32 / RTOS
Communication/NetworkCommunication / 5G / Data Communication / Protocols / RF
TestingTesting / Test Development / Automated Testing / QA

流程

Process

  1. 解析简历：提取目标岗位、技能列表、2–5 个项目（名称、技术栈、一句话描述、亮点/难点）。信息缺失则留空，禁止编造。
  2. 检索参考面经（按优先级 Grep / Read）：
    • 先：
      knowledge/experiences/platform/
      （CodeFun2000 站内面经）
    • 再：
      knowledge/wiki/compiled/
      、
      knowledge/interview/
    • 后：
      knowledge/experiences/hw-exp-*-ext-*.md
      、
      hw-exp-*-xhs-*.md
    • 按岗位关键词与技能匹配，选出 3–4 条最相关面经用于「参考面经」列表
  3. 选题手撕（Agent 内部）：
    • 仅从 
      knowledge/coding-problems/hot100/index.json
      选题并取可点击 
      source_url
      / 
      codefun_url
      / 
      leetcode_url
    • 需要题意时再 Read 对应题面；禁止用第三方求职社区或库外 URL 充当手撕链接
    • 推荐理由须结合简历技能或项目，2–3 句
  4. 生成面试题：
    • 项目提问 3 条：必须点名简历中的具体项目与技术细节，忌空泛
    • 八股文考点 3 条：贴合技能列表，可与项目结合
    • 手撕题 1 道：来自上一步索引
  5. 输出严格使用下方 Markdown 结构；可一次给出 1 组；若用户要求多组，最多再给 2 组变体（换角度，勿重复同一项目问法）
  1. Resume Parsing: Extract the target job, skill list, 2–5 projects (name, tech stack, one-sentence description, highlights/difficulties). Leave blank if information is missing, prohibit fabrication.
  2. Retrieve Reference Interview Experiences (priority: Grep / Read):
    • First: 
      knowledge/experiences/platform/
      (CodeFun2000 internal interview experiences)
    • Then: 
      knowledge/wiki/compiled/
      , 
      knowledge/interview/
    • Last: 
      knowledge/experiences/hw-exp-*-ext-*.md
      , 
      hw-exp-*-xhs-*.md
    • Match by job keywords and skills, select 3–4 most relevant interview experiences for the "Reference Interview Experiences" list
  3. Select Coding Problems (Internal to Agent):
    • Only select from 
      knowledge/coding-problems/hot100/index.json
      and retrieve clickable 
      source_url
      / 
      codefun_url
      / 
      leetcode_url
    • Read the corresponding problem description only when needed; prohibit using third-party job hunting communities or external URLs as coding problem links
    • The recommendation reason must be combined with resume skills or projects, 2–3 sentences
  4. Generate Interview Questions:
    • Project Questions: 3 questions: Must specify specific projects and technical details in the resume, avoid vague questions
    • Theoretical Knowledge Points: 3 questions: Align with the skill list, can be combined with projects
    • Coding Problem: 1 question: From the index in the previous step
  5. Output strictly follows the Markdown structure below; can provide 1 set at a time; if the user requests multiple sets, provide up to 2 more variants (change perspective, do not repeat the same project questioning method)

方案角度（轮换选用）

Solution Perspectives (Rotate Selection)

  • 从项目技术选型和难点切入
  • 从项目数据量、性能优化和实际效果切入
  • 从团队协作、遇到的问题和解决方案切入
  • 从八股基础与项目结合点切入
  • 从业务场景和算法/架构选择切入
  • 从扩展性、可维护性和工程落地切入
  • Start from project technical selection and difficulties
  • Start from project data volume, performance optimization and actual effects
  • Start from team collaboration, problems encountered and solutions
  • Start from the combination of basic theoretical knowledge and projects
  • Start from business scenarios and algorithm/architecture selection
  • Start from scalability, maintainability and engineering implementation

输出格式

Output Format

markdown
undefined
markdown
undefined

简历模拟面试（{岗位方向中文}）

Resume Mock Interview ({Chinese Job Direction})

方案角度：{角度一句话}
Solution Perspective: {one-sentence perspective}

项目提问

Project Questions

  1. …
  2. …
  3. …
  1. …
  2. …
  3. …

八股文考点

Theoretical Knowledge Points

  1. …
  2. …
  3. …
  1. …
  2. …
  3. …

手撕题

Coding Problem

题目标题 — 简短考点提示
推荐理由：结合简历中的具体技能或项目说明为何选这题。
参考面经：
  • [B] 站内面经标题 — CodeFun2000
  • [C] 外部面经《…》
  • [C] 小红书面经《…》
undefined
[Problem Title](clickable link) — Brief knowledge point hint
Recommendation Reason: Explain why this question is selected by combining specific skills or projects in the resume.
Reference Interview Experiences:
undefined

参考面经引用规则（同 
/hw-ask
）

Reference Interview Experience Citation Rules (Same as 
/hw-ask
)

  • CodeFun2000 / 自有：带链接；外部面经/小红书：只写「外部面经」或「小红书面经」+ 标题，禁止外链，外部面经禁止写原站点名/域名
  • B 站若引用：须 Read 片段并核对相关度，格式 
    [A] 塔子哥公开课 […](url?t=秒) — B站 **MM:SS–MM:SS**
    ；禁止写「字幕」
  • 禁止在输出中出现 
    knowledge/
    、
    index.json
    、
    hot100
    、卡片 id、Grep/Read、「知识库」等内部用语
  • CodeFun2000 / Internal: with link; External interview experiences/Xiaohongshu: only write "External Interview Experience" or "Xiaohongshu Interview Experience" + title, prohibit external links, do not write the original site name/domain for external interview experiences
  • If citing Bilibili: Must read the segment and check relevance, format as 
    [A] Tazige Public Course […](url?t=seconds) — Bilibili **MM:SS–MM:SS**
    ; prohibit writing "subtitles"
  • Prohibit internal terms such as 
    knowledge/
    , 
    index.json
    , 
    hot100
    , card ID, Grep/Read, "knowledge base" in the output

禁止

Prohibited Actions

  • 编造简历中不存在的项目或技术
  • 声称精准押题或给出虚假匹配百分比
  • 手撕题链接不在 
    hot100/index.json
    内
  • 输出小红书 URL 或任何第三方求职社区域名/外链
  • 向客户暴露仓库路径或检索过程
  • 在未拿到简历时直接生成完整题单
  • Fabricate projects or technologies not present in the resume
  • Claim accurate question prediction or provide false matching percentages
  • Use coding problem links not in 
    hot100/index.json
  • Output Xiaohongshu URLs or any third-party job hunting community domain names/external links
  • Expose repository paths or retrieval processes to customers
  • Generate a complete question list directly without obtaining the resume first

相关

Related

流程/政策答疑请改用 
/hw-ask
。性格测评练习见 
/hw-ask
或 测评模拟。
For process/policy questions, please use 
/hw-ask
 instead. For personality test practice, see 
/hw-ask
or Assessment Simulation.