Internship Project Preparation Tool
LiuMengxuan04/shushu-internship-tool
Turns a target internship job description into a resume-ready, interview-ready project by finding and auditing GitHub projects and drafting resume bullets and interview Q&A.
Builds a personalized Chinese interview prep packet by matching a resume against real interview-experience posts, deduped and ranked by recency.
SKILL.md written in Chinese; this summary is our English description.
Install Resume-Matched Interview Prep Builder
$ npx skills add KunChen1110/InterviewRadar --skill interview-radar -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install KunChen1110/InterviewRadar interview-radar --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
Claude Code skills documentation · loads skills from .claude/skills/
Install the "interview-radar" agent skill from https://github.com/KunChen1110/InterviewRadar/tree/main into .claude/skills/interview-radar/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "interview-radar", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add KunChen1110/InterviewRadar --skill interview-radar -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install KunChen1110/InterviewRadar interview-radar --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "interview-radar" agent skill from https://github.com/KunChen1110/InterviewRadar/tree/main into .agents/skills/interview-radar/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "interview-radar", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add KunChen1110/InterviewRadar --skill interview-radar -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install KunChen1110/InterviewRadar interview-radar --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "interview-radar" agent skill from https://github.com/KunChen1110/InterviewRadar/tree/main into .cursor/skills/interview-radar/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "interview-radar", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add KunChen1110/InterviewRadar --skill interview-radar -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install KunChen1110/InterviewRadar interview-radar --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "interview-radar" agent skill from https://github.com/KunChen1110/InterviewRadar/tree/main into .gemini/skills/interview-radar/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "interview-radar", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install KunChen1110/InterviewRadar interview-radarInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add KunChen1110/InterviewRadar --skill interview-radar -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "interview-radar" agent skill from https://github.com/KunChen1110/InterviewRadar/tree/main into .github/skills/interview-radar/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "interview-radar", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add KunChen1110/InterviewRadar --skill interview-radar -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install KunChen1110/InterviewRadar interview-radar --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "interview-radar" agent skill from https://github.com/KunChen1110/InterviewRadar/tree/main into .opencode/skills/interview-radar/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "interview-radar", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
interview-radar- GitHub stars
- 248
- Token cost
- ~2.6k tokens
- SKILL.md length
- 455 words
- Files
- 91 (incl. scripts, assets)
- Skills in repo
- 1
- Repo updated
- First seen
- Licence
- MIT
At a glance
Builds a personalized Chinese interview prep packet by matching a resume against real interview-experience posts, deduped and ranked by recency.
- Works in 6 steps: 准备(仅当启用小红书源)。 两种模式二选一,见… → 简历理解。 调用 extract_resume。若 needs_vision… → 种子查询生成。 用你自己的领域知识,从「岗位方向 +… → …
- Preparing for an interview in a loosely defined role direction
- SKILL.md covers 输入, 工具(用包内 venv 运行:.venv/bin/python), 工作流 and 备考包固定模板, plus 1 more section
- Calls python
What it does
This skill extracts text from a resume in PDF, image, or plain-text form, then searches a vague job direction across GitHub, Nowcoder, and Xiaohongshu connectors, keeping only posts from the last two years, deduplicating them, and ranking by frequency and recency before pulling project-specific follow-up questions tied back to the resume.
Xiaohongshu content can come from a pre-exported JSON file or from a bundled scraping driver using QR-code or cookie login, in either a fast mode that reads titles, captions, and tags, or a deep mode that downloads and runs OCR on images as the main text. The agent handles judgment calls while bundled Python scripts handle the deterministic extraction, scraping, and ranking, communicating through cached JSON files.
When your agent uses it
- Preparing for an interview in a loosely defined role direction
- Building a resume-matched question bank from real interview experiences
- Collecting recent interview posts for a target role
Example prompts
- “Build my interview prep packet for an AI application developer role from my resume.”
- “Find recent interview posts for a marketing internship.”
- “Generate project follow-up questions based on my resume and this job direction.”
Requirements
- Python virtual environment bundled with the skill
- A scraping driver installation for Xiaohongshu (optional)
Workflow steps
6 steps, taken from the first numbered list in SKILL.md.
- 准备(仅当启用小红书源)。 两种模式二选一,见 docs/setup/mediacrawler.md
- 简历理解。 调用 extract_resume。若 needs_vision 为真,就用你自己的视觉能力直接读这张图片/PDF。产出结构化摘要:技能、项目(每个项目用到的技术)、关键术语。
- 种子查询生成。 用你自己的领域知识,从「岗位方向 + 简历」推导出种子查询。这是领域无关的:无论什么领域(市场、量化、后端、设计……)你本来就知道相关的岗位别名和底层技能/话题,当场生成即可。不要依赖任何预设词表。 种子来自两处:(a) 岗位方向隐含的相关岗位别名;(b)…
- 迭代检索。 源的优先级:牛客 + 小红书(主力,带时间戳)> GitHub(补充,常过时)。
- 内容级相关性判定。 通过读帖子内容对照用户岗位 + 简历来判断每条是否相关——不是靠帖子的岗位名是否匹配某张预设表。
- 候选抽题与语义决策。 文本类 RawPost 直接使用 content_text/raw_text。图片类 RawPost(小红书,post_type="image"):正式包默认用 deep 模式读 content_text/raw_text,它优先来自图片…
What it can do on your machine
Read from SKILL.md and the folder at commit f551c9c. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships 1 file in scripts/, which the agent can run.
Shell commands in SKILL.md call:
python
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Resume-Matched Interview Prep Builder loads about 2.6k tokens when it runs. Until then it costs about 44 tokens; SKILL.md has 455 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
SKILL.md
The full file from KunChen1110/InterviewRadar at commit f551c9c, republished under its MIT licence (© KunChen1110). 455 words, ~2,568 tokens.
Download SKILL.mdSave it as
.claude/skills/interview-radar/SKILL.md (or your agent's skills folder). This skill also uses 90 other files; get the full folder from GitHub.- name
- interview-radar
- description
- 面试准备 / interview prep。当用户上传简历(PDF/图片)并给出一个模糊的目标岗位方向(例如"AI 应用开发""市场实习")时使用。广撒网检索真实面经(主力源:牛客 + 小红书;补充源:GitHub + 通用正文页),只保留近两年内容,去重并按频次+时效排序,结合用户简历生成带项目追问的个性化备考包。
InterviewRadar · 面试雷达 Skill
把「简历 + 模糊岗位」变成一份基于真实面经内容的个性化备考包。你(agent)负责推理判断;scripts/ 下的 Python 脚本负责确定性的脏活。两者通过 corpus_cache/ 里的 JSON 文件交互。
输入
- 简历:PDF、图片或文本文件路径。
- 模糊岗位:一个方向,例如"AI 应用开发"(不是具体的 JD)。
工具(用包内 venv 运行:.venv/bin/python)
scripts/resume_extract.py→extract_resume(path) -> ResumeExtraction{text, needs_vision, asset_path}scripts/connectors/github.py→GithubConnector(repo_raw_urls).search(queries) -> SearchResultscripts/connectors/nowcoder.py→NowCoderConnector(post_urls).search(queries) -> SearchResultscripts/connectors/xiaohongshu.py→XiaohongshuConnector(export_path=..., driver=..., enable_image_ocr=...).search(queries) -> SearchResult(二选一:export_path读预生成的 JSON;driver自动跑 MediaCrawler;enable_image_ocr=True为 deep 模式,图片会下载/OCR 后作为主正文;False为 fast 模式,只读标题、正文 caption、标签和时间戳)scripts/scrape/mediacrawler_driver.py→MediaCrawlerDriver(home=None).scrape_xhs(keywords, login_type="qrcode"|"cookie") -> Path,驱动模式——shell out 调本机已装的 MediaCrawler(默认从$MEDIACRAWLER_HOME或~/.mediacrawler/找)。推荐 cookie 登录:在 MediaCrawlerconfig/base_config.py里设置LOGIN_TYPE = "cookie"和COOKIES = "web_session=<value>"。scripts/scrape/normalize_xhs.py→normalize(notes) -> list[dict](CLI:python -m scripts.scrape.normalize_xhs <in.json> -o <out.json>),把 MediaCrawler 原生输出归一化为XiaohongshuConnector的输入。仅手动模式用得到;driver 模式连接器内部自动调用。scripts/ocr/extract.py→extract_text_from_image(path, engine=None, min_confidence=0.6) -> OcrResult{text, confidence, needs_vision}scripts/ocr/xhs_images.py→ 下载小红书image_list,默认探测 RapidOCR,把分页 OCR 合并进RawPost.content_text/raw_text;低质量时设needs_vision_fallback=Truescripts/corpus/extraction.py→extract_candidates(raw_posts) -> list[ExtractionCandidate]+materialize_questions(candidates, decisions) -> ExtractionMaterializationResult(候选原文切分 + 受约束决策落题)scripts/corpus/store.py→save_raw_posts / load_raw_posts / save_questions / load_questions / save_extraction_candidates / load_extraction_candidates / save_extraction_decisions / load_extraction_decisionsscripts/corpus/runner.py→run_pipeline(raw_posts_path, decisions_path, output_dir, today=None) -> PipelineRunResult(离线、可复放地生成输入副本、候选、落题、排序题、拒绝项、诊断和 SHA-256 manifest)scripts/corpus/recency.py→classify_recency(...)+filter_recent(posts, window_days=730, today=None, keep_undated=True) -> list[RawPost]scripts/corpus/dedupe_rank.py→dedupe_and_rank(questions) -> list[Question](先按独立来源数,再按出现次数和时效排序;question_rank_score()可查看评分明细)scripts/corpus/quality.py→prepare_questions(raw_posts, questions, ...) -> QuestionPreparationResult(证据校验、时效过滤、去重排序的统一入口;新运行默认拒绝无证据题目)- 数据模型在
scripts/models.py;结构说明见assets/schema.md。
工作流
准备(仅当启用小红书源)。 两种模式二选一,见
docs/setup/mediacrawler.md:- driver 模式(推荐):用户一次性装 MediaCrawler 并登录。优先用 cookie 模式:从正常浏览器复制
web_session,写入 MediaCrawlerconfig.COOKIES,然后用XiaohongshuConnector(driver=MediaCrawlerDriver(), login_type="cookie")自动跑采集。二维码模式仍可用,但更容易触发风控。 - 手动模式:用户每次自己跑 MediaCrawler +
normalize_xhs.py,把corpus_cache/xhs_export.json喂给XiaohongshuConnector(export_path=...)。 - 读取深度必须说明:小红书很多面经受文字区限制,完整题目在图片里。
fast模式只读标题/正文/标签,适合验证召回;正式备考包优先用deep模式(enable_image_ocr=True)读取图片 OCR。若因速度或依赖问题使用fast,必须在输出中明确写“未读取图片 OCR,可能漏掉图片里的完整题目”。 文本/牛客/GitHub 源不需要这一步。
- driver 模式(推荐):用户一次性装 MediaCrawler 并登录。优先用 cookie 模式:从正常浏览器复制
简历理解。 调用
extract_resume。若needs_vision为真,就用你自己的视觉能力直接读这张图片/PDF。产出结构化摘要:技能、项目(每个项目用到的技术)、关键术语。种子查询生成。 用你自己的领域知识,从「岗位方向 + 简历」推导出种子查询。这是领域无关的:无论什么领域(市场、量化、后端、设计……)你本来就知道相关的岗位别名和底层技能/话题,当场生成即可。不要依赖任何预设词表。 种子来自两处:(a) 岗位方向隐含的相关岗位别名;(b) 从简历里抽出的具体技能/项目/关键词。优先用底层技能/话题词,而不是岗位名——它们更稳定、召回更好。
迭代检索。 源的优先级:牛客 + 小红书(主力,带时间戳)> GitHub(补充,常过时)。
3a. URL 发现(每轮先做)。 用你的搜索能力(WebSearch 或等价工具)对当前的种子查询跑一遍,收集候选 URL。按域名分桶:
nowcoder.com/discuss/<post_id>→ 进NowCoderConnector(post_urls=...)xiaohongshu.com/explore/<note_id>→ 不能直抓;如果启用了小红书源,把当前的关键词丢给XiaohongshuConnector(driver=MediaCrawlerDriver()).search([keywords]),它会自动 shell out 跑 MediaCrawler。若未启用,记下笔记 ID 让用户按docs/setup/mediacrawler.md配置github.com/<owner>/<repo>/blob/<branch>/<path>或raw.githubusercontent.com/...→ 转 raw URL → 进GithubConnector(repo_raw_urls=...)- 其他公开正文页(知乎 article、CSDN 文章、个人博客、woshipm/uisdc 等):用 WebFetch 拉回正文,自己手工构造
RawPost(source="webfetch:<domain>", post_type="text", raw_text=<正文>, posted_at=<页面可见日期或 None>),和 connector 结果一起save_raw_posts
显式排除:聚合/listing 页(只接 article 页);403/需登录的页面(不浪费 fetch 配额,记下来当作知识缺口告诉用户)。
3b. 调 connectors + 收割。 把分好桶的 URL 喂给对应 connector,结果用
save_raw_posts落盘。读取结果,收割真实出现的岗位名 / 标签 / 高频术语,用收割到的词跑下一轮 3a,直到不再冒出新词。若某 connector 返回status="degraded"(例如牛客需要 cookie、小红书需要先跑 MediaCrawler、或消息含selector表示 HTML 漂移),把它需要的东西告诉用户;主力源降级会显著影响时效性,必须明确提示用户,不要默默用 GitHub 凑数。GitHub 调用必带
relevance_hints(不要传None或[])。 GitHub 仓库里夹带大量算法/八股,不过滤会严重污染语料。GithubConnector(repo_raw_urls=..., relevance_hints=<当前一轮的术语/岗位别名>)。冷启动(第一轮还没有收割结果)时,直接把步骤 2 的种子查询当 hints 传进去——任何时候都要传非空 list。命中规则:子串、大小写不敏感,只要正文里出现任一 hint 就保留。Human-in-the-loop: 在最后一轮之前,把你从真实数据里发现的方向/术语展示给用户,让他增删/纠偏。
内容级相关性判定。 通过读帖子内容对照用户岗位 + 简历来判断每条是否相关——不是靠帖子的岗位名是否匹配某张预设表。
候选抽题与语义决策。 文本类 RawPost 直接使用
content_text/raw_text。图片类 RawPost(小红书,post_type="image"):正式包默认用 deep 模式读content_text/raw_text,它优先来自图片 OCR;标题、caption、tags 只在locator_text里做召回/定位。若needs_vision_fallback=True,先按asset_paths顺序补读图片,把视觉补读的逐字原文写入该帖content_text(保留已有 OCR 文本),再继续抽题。若本次为了速度用了 fast 模式(enable_image_ocr=False),只能把标题/正文/标签作为线索,不得声称覆盖了图片里的完整面经题。对所有
RawPost调用extract_candidates(raw_posts),并用save_extraction_candidates写入corpus_cache/extraction_candidates.json。候选只是原文片段,不是题目。读取候选后,结合目标岗位和简历,每个候选写一条ExtractionDecision(candidate_id, accepted, canonical_text, topic, role_tags)并用save_extraction_decisions写入corpus_cache/extraction_decisions.json:accepted=true只用于真实、相关且语义明确的面试问题;canonical_text用一句尽量短的中文概括问题意图,去掉“请介绍/如何看待/面试官问”等话术,但必须保留技术实体、业务对象和关键条件。例如“如何评估招聘渠道”和“招聘渠道应该怎么评估”都写成“评估招聘渠道”;“Redis 缓存穿透”和“Redis 缓存雪崩”不能合并。- 不要提供题面、URL、日期或原文证据。它们由 Python 从候选固定生成;拒绝的候选也必须写
accepted=false,不得静默跳过。
Show full SKILL.md (127 more words)Show less
5a. 生成可复放运行包。 决策写完后，执行：
python -m scripts.corpus.runner \
--raw-posts corpus_cache/raw_posts.json \
--decisions corpus_cache/extraction_decisions.json \
--output corpus_cache/runs/<本次标识> \
--today YYYY-MM-DD --output 必须是一个尚不存在的目录。runner 会重新从 raw_posts 生成候选，再完成 materialize_questions()、prepare_questions() 和去重排序；它把输入副本、中间产物、排序题、拒绝项、诊断和 SHA-256 manifest 一起写入该目录。manifest 还会记录实际用于时效判断的 today;正式复跑应显式传入同一日期。最终题面仍来自候选的 display_text,原文证据来自候选的 excerpt,不得手工替换。
5b. 读取数据缺口。 先检查 <输出目录>/diagnostics.json:其中包含 decision 处理结果（包括明确不保留或漏决策的候选）、候选/证据问题、过期来源计数和被拒绝项。它不执行搜索、爬虫、简历阅读、语义 decision 或最终备考文案；这些仍是前后显式步骤。正式备考包只允许读取 <输出目录>/ranked_questions.json;不得回退到 materialized_questions.json、旧缓存或人工拼接题目。
去重 & 排序。 runner 内部的
prepare_questions()已完成证据与时效校验、按canonical_text去重和排序。排序优先考虑独立来源 URL 数和时效,出现次数只用于同分时破平。同一 URL 重复出现不能冒充多来源高频。只有至少 2 个经过 evidence 校验的独立来源支持的问题才称为“高频”;其余可作为“相关题”保留,但必须明确是单来源。项目锚定推理。 对每道高频题,检查它能否挂到简历里的某个项目/技能上。能挂上就构造
FollowUpChain(种子题 → 个性化追问,is_grounded=true)。每个追问都必须能追溯到(简历某项目/技能)+(某条真实爬到的题);追溯不上就设is_grounded=false,当普通八股题保留。不要凭空编追问。备考包。 严格按下面的固定模板写 Markdown,保存到
corpus_cache/prep_package.md并展示给用户。不要自由发挥改大结构;若某部分数据不足,写清"数据不足/未覆盖",不要编。
备考包固定模板
# {目标岗位}岗位备考包 — {候选人姓名或"候选人"}
生成日期:{YYYY-MM-DD}
目标岗位:{用户给定岗位 + 从真实数据收割到的岗位别名}
## 1. 你的候选人定位
用 1 段话定义候选人的面试包装方向。格式:
> {一句候选人定位}
简历里最强的三个证据:
1. **{项目/经历 1}**
- {证据点 1}
- {证据点 2}
- {适合回答什么面试问题}
2. **{项目/经历 2}**
- ...
3. **{项目/经历 3}**
- ...
## 2. 岗位 Gap 分析
| 维度 | 当前简历表现 | 面试风险 | 准备建议 |
|---|---|---|---|
| {能力维度} | {基于简历的证据} | {真实面经会追问的风险} | {可执行补强建议} |
至少覆盖:技术理解、产品基本功、业务 sense、数据/实验、落地可信度。非 AI 岗位时把维度换成该岗位的核心能力。
## 3. 真实数据来源概况
本次召回:
- {来源 1}:{数量/状态/保存路径}
- {来源 2}:{数量/状态/保存路径}
数据缺口:
- {降级源、OCR/反爬/时间戳缺口等}
## 4. 高频题 Top {N}
### {序号}. {题目}
来源:
- {source title 或 source type}:`{url}`
- 原文证据:`{QuestionEvidence.excerpt}`
- ...
回答要点/回答框架:
- {要点 1}
- {要点 2}
- {要点 3}
可挂简历锚点:
> {把这题连接到候选人某个项目/技能的一段话}
如果适合表格解释,可以用小表格;否则用 bullet。每题必须至少有 1 个真实来源。没有来源的题不能进 Top 高频题。
## 5. 个性化项目追问链
### 链 {序号}:{主题} → {简历项目/经历}
种子题:{来自第 4 节的高频题}
追问:
1. {追问 1}
2. {追问 2}
3. {追问 3}
4. {追问 4}
5. {追问 5}
准备重点:
- {如何准备真实例子/图/指标}
- {面试时要强调的产品视角}
每条追问必须同时能追溯到"真实面经题目"和"简历项目/技能";追溯不上就不要写成个性化追问。
## 6. 你的 60-90 秒自我介绍草稿
写 2-3 段中文口语稿。必须包含:
- 背景
- 目标岗位动机
- 2-3 个最强项目证据
- 候选人的差异化定位
## 7. 一周冲刺计划
### Day 1:{主题}
- {行动项}
- {行动项}
...
### Day 7:{主题}
- {行动项}
- {行动项}
## 8. 建议你立刻补强的简历表述
### {项目/经历 1}
补:
> {可直接放进简历/面试话术的改写}
### {项目/经历 2}
补:
> ...
## 9. 来源列表
按来源类型分组列出代表来源:
小红书:
- `{url}` — {一句说明}
牛客/网页/GitHub:
- `{url}` — {一句说明}
## 10. 面试前速查清单
- 60 秒自我介绍
- {3 个最能证明岗位匹配度的项目/经历证据}
- {3 个真实 trade-off / 失败 / 返工案例}
- {5 个目标岗位必会高频题;例如 AI 产品岗可写 RAG、Agent、实验、指标、产品设计,其他岗位按真实面经替换}
- {4 个高质量反问}
- {可以展示的作品、流程图、数据、文档或代码证据}模板约束:
- 全文中文,但技术名词可保留英文。
- 第 4 节的题目只能来自本次运行包的
<输出目录>/ranked_questions.json;它的每条QuestionEvidence.excerpt必须可追溯到本次运行包中的raw_posts.json。小红书 caption/tag 只作为locator_text辅助,不能单独生成题目。 - 不得直接向模板输入任意
Question、RawPost.content_text/raw_text或手工调用prepare_questions()的结果来绕过运行包。只有 URL、没有原文片段的题不能进入高频区。 - 高频题优先按
dedupe_and_rank结果排序;人工调整时只能因岗位相关性或简历匹配度调整,并说明依据。称为“高频”的题必须至少有 2 个独立来源 URL;不足时改写为“相关题 Top N”。 - 来源 URL 必须真实存在于
source_refs或 RawPost.url;不要写"综合资料"这类不可追溯来源。 - 自我介绍和简历表述可以做表达优化,但事实必须来自简历或真实面经。
- 第 10 节必须按用户目标岗位动态生成,不要写死为 AI 产品题;AI 产品只是示例。
约束
- 所有面向用户的产出一律用中文(备考包、题目、追问、分析)——面经源是中文。
- 当前源:牛客 + 小红书(主力,带时间戳)+ GitHub(补充)。
- 小红书走 MediaCrawler 采集导出(用户预先离线跑一次,流程见
docs/setup/mediacrawler.md),OCR 采用混合策略(粗 OCR + 视觉回退);MediaCrawler 仅供个人、非商业用途。 - 时效性是硬需求:有明确日期的来源只保留近两年;无日期补充源必须单独披露并按最低权重排序。
- 后续版本用到的第三方爬虫(如 MediaCrawler)仅供个人、非商业用途。
- 可追溯优先于流畅度:绝不编造无法追溯到真实数据的题目或追问。
- 若 connector 返回
degraded且消息含selector,说明源站点 HTML 改了选择器;到对应scripts/connectors/<name>.py顶部注释看当前假设,核对真实 HTML 后更新选择器并补 fixture。
© KunChen1110, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 90 other files (scripts, assets) in the repository root of KunChen1110/InterviewRadar.
- SKILL.md
- .github/workflows/tests.yml
- .gitignore
- DISCLAIMER.md
- LICENSE
- README.md
- assets/logo.png
- assets/schema.md
- assets/workflow.png
- docs/authorized-source-intake.md
- docs/plans/2026-05-28-interview-intelligence-mvp.md
- docs/plans/2026-05-28-interview-intelligence-nowcoder-recency.md
- docs/plans/2026-05-29-interview-intelligence-xiaohongshu-ocr.md
- docs/plans/2026-05-31-mediacrawler-adapter.md
- docs/plans/2026-06-01-github-hints-and-undated-demote.md
- docs/plans/2026-06-01-nowcoder-selector-fix.md
- … and 75 more
Open the folder on GitHubat commit f551c9c
Compare with similar skills
Resume-Matched Interview Prep Builder next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Resume-Matched Interview Prep Builder this skillKunChen1110/InterviewRadar | 248 | — | ~2.6k | Automated safety check: Pass | MIT | |
| Internship Project Preparation ToolLiuMengxuan04/shushu-internship-tool | 2.1k | — | ~2.3k | Automated safety check: Pass | Custom licence | |
| Backend and Agent Project Selectorlishuangqiang/backend-agent-resume-scout | 347 | — | ~1.4k | Automated safety check: Pass | Apache-2.0 | |
| Internship-to-Resume ToolkitSunanzhe2004/shushu-internship-resume-optimizer | 123 | — | ~1.6k | Automated safety check: Pass | Custom licence | |
| Leetcode Pywislertt/leetcode-py | 142 | — | ~1.4k | Automated safety check: Pass | Apache-2.0 | |
| Python Backend Interview QuestionsSnailclimb/interview-guide | 3.3k | — | ~119 | Automated safety check: Pass | AGPL-3.0 |
Similar skills
Backend and Agent Project Selector
lishuangqiang/backend-agent-resume-scout
Finds backend or AI agent projects on GitHub that are worth putting on a resume, checks them against local source and writes a Markdown resume package.
Internship-to-Resume Toolkit
Sunanzhe2004/shushu-internship-resume-optimizer
Turns internship code repos, project summaries and business documents into resume-ready bullets, JD-matched rankings, and interview-ready STAR material.
Leetcode Py
wislertt/leetcode-py
Generates Python LeetCode practice environments and manages a 307-problem catalog with the lcpy CLI.
Python Backend Interview Questions
Snailclimb/interview-guide
Acts as a Python backend interviewer, setting practical questions on Python basics, databases, Django and Flask, caching and deployment, with follow-ups on real production work.
Homepage Generator
wanshuiyin/ARIS-in-AI-Offer
Generate a fact-checked academic personal homepage from a CV, optionally augmented by an existing manual homepage and an assets directory.
Related
Works with
Categories
Questions about Resume-Matched Interview Prep Builder
What does Resume-Matched Interview Prep Builder do?
Builds a personalized Chinese interview prep packet by matching a resume against real interview-experience posts, deduped and ranked by recency. This skill extracts text from a resume in PDF, image, or plain-text form, then searches a vague job direction across GitHub, Nowcoder, and Xiaohongshu connectors, keeping only posts from the last two years, deduplicating them, and ranking by frequency and recency before pulling project-specific follow-up questions tied back to the resume.
When should I use Resume-Matched Interview Prep Builder?
Resume-Matched Interview Prep Builder fits situations like: preparing for an interview in a loosely defined role direction; building a resume-matched question bank from real interview experiences; collecting recent interview posts for a target role.
How do I install Resume-Matched Interview Prep Builder in Claude Code?
Run `npx skills add KunChen1110/InterviewRadar --skill interview-radar -a claude-code`. Or copy the skill folder (the KunChen1110/InterviewRadar repository) into .claude/skills/interview-radar in your project. Claude Code loads it when a task matches its description.
How do I install Resume-Matched Interview Prep Builder in Codex?
Run `npx skills add KunChen1110/InterviewRadar --skill interview-radar -a codex`. Or copy the skill folder (the KunChen1110/InterviewRadar repository) into .agents/skills/interview-radar in your project. Codex loads it when a task matches its description.
Can I use Resume-Matched Interview Prep Builder in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add KunChen1110/InterviewRadar --skill interview-radar -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/interview-radar, .gemini/skills/interview-radar, .github/skills/interview-radar and .opencode/skills/interview-radar in your project.
What does Resume-Matched Interview Prep Builder need to run?
Going by SKILL.md and its folder, Resume-Matched Interview Prep Builder needs the command-line tools its instructions call (python). Our summary lists: Python virtual environment bundled with the skill; A scraping driver installation for Xiaohongshu (optional).
Does Resume-Matched Interview Prep Builder access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Resume-Matched Interview Prep Builder safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
What licence does Resume-Matched Interview Prep Builder use?
Resume-Matched Interview Prep Builder is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Resume-Matched Interview Prep Builder use?
About 2.6k tokens (SKILL.md is roughly 10k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Resume-Matched Interview Prep Builder?
Skills that share tags, products or a category with Resume-Matched Interview Prep Builder: Internship Project Preparation Tool (LiuMengxuan04/shushu-internship-tool, 2.1k stars), Backend and Agent Project Selector (lishuangqiang/backend-agent-resume-scout, 347 stars), Internship-to-Resume Toolkit (Sunanzhe2004/shushu-internship-resume-optimizer, 123 stars) and Leetcode Py (wislertt/leetcode-py, 142 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Resume-Matched Interview Prep Builder?
KunChen1110 (a GitHub user) maintains it in KunChen1110/InterviewRadar, which has 248 GitHub stars. The repository was last updated on August 20, 2026.
Source: KunChen1110/InterviewRadar on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.