agentleFS
lavine888lavine888/career-alpha/skills/interview/SKILL.md

interview

Skill0 starsChanged 3 months agoMITView on GitHub

Career Alpha 面试防守技能。用于根据简历、项目和 Claim–Evidence Ledger 连续追问,检验候选人是否真的理解自己写下的经历;当用户要模拟面试、预测追问、检查项目是否讲得住、识别 AI 包装痕迹、准备行为/技术面,或决定某条简历主张该保留还是降级时使用。

94 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use fetch-library-docs, browsing-with-playwright and 8 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Rehan-Ul-HaqSkill

    fetch-library-docs

    Fetches official documentation for external libraries and frameworks (React, Next.js, Prisma, FastAPI, Express, Tailwind, MongoDB, etc.) with 60-90% token savings via content-type filtering. Use this skill when implementing features using library APIs, debugging library-specific errors, troubleshooting configuration issues, installing or setting up frameworks, integrating third-party packages, upgrading between library versions, or looking up correct API patterns and best practices. Triggers automatically during coding work - fetch docs before writing library code to get correct patterns, not after guessing wrong.

    Also in 4 projects that use this

  • mjunaidcaSkill

    browsing-with-playwright

    Browser automation using Playwright MCP. Navigate websites, fill forms, click elements, take screenshots, and extract data. Use when tasks require web browsing, form submission, web scraping, UI testing, or any browser interaction. NOT when only fetching static content (use curl/wget instead).

    Also in 4 projects that use this

  • qa-amanSkill

    teach-me

    Teach a topic, video, article, codebase change, spec, or transcript until the user demonstrably understands it, using staged instruction, a running checklist doc, and quizzes. Use whenever the user says "teach me", "explain this in detail", "help me understand X", "/teach-me", "walk me through this", "I want to learn X", "understand this deeply", or pastes a YouTube URL, article, paper, PR, or spec and asks to understand it rather than summarize it. Also use when the user wants study notes or raw material from a source they are learning. Do NOT use for a quick factual lookup, a summary request, or when the user just wants the answer.

    Also in 8 projects that use this

  • trebaudSkill

    find-unknowns

    Workflow for surfacing a task's unknowns before, during, and after implementation — blind spot pass, brainstorm/prototype, interview, references, implementation plan, deviation notes, quiz. Use when starting work in an unfamiliar domain or codebase, when a task feels underspecified, when the user says "find my unknowns", "blindspot pass", "what am I missing", "help me scope this", or when a long task came back wrong.

    Also in 3 projects that use this

  • galsapirSkill

    long-horizon

    Transform a user's draft prompt into a long-horizon coding-agent task prompt. Use when the user asks to make a prompt long-horizon, turn something into a long-running agent task, add long-horizon scaffolding, or invokes "$long-horizon" or "/long-horizon". Appends framing for running far without unnecessary stops, using subagents for scoped background work, and maintaining implementation-notes.html with decisions, deviations, tradeoffs, and open questions.

    Also in 3 projects that use this

  • muhammad-anas35Skill

    skill-creator-pro

    Creates production-grade, reusable skills that extend Claude's capabilities. This skill should be used when users want to create a new skill, improve an existing skill, or build domain-specific intelligence. Gathers context from codebase, conversation, and authentic sources before creating adaptable skills.

    Also in 4 projects that use this

The file

Files
SKILL.md

/interview — Defend Your Claims

/interview 的目标不是陪用户背答案,而是找到:

简历上哪一句最容易被问穿,为什么会被问穿,以及应该补知识、补证据还是直接降级表述。

它默认站在一个认真、怀疑、但公平的面试官视角。


1. Inputs

优先读取:

  • /proof Claim–Evidence Ledger;
  • /position Safe / Strong / Future Position;
  • 当前简历;
  • 目标 JD;
  • 公司类型;
  • 面试阶段;
  • 用户想练的语言 /形式。

如果没有 ledger,可以先从简历里抽取高风险 claim,但要提醒:未审计事实的 risk judgment 可能不完整。


2. Build an Interview Risk Map

优先攻击高风险内容:

  • 精确数字;
  • led / owned / architected / drove;
  • production / large-scale;
  • 复杂 architecture;
  • benchmark;
  • 比赛名次;
  • 用户量 / revenue / performance;
  • open-source contribution;
  • AI-assisted coding;
  • 团队成果与个人边界。

输出:

ClaimEvidence confidenceInterview riskWhy it will be challengedFirst question

3. Use claim-specific question ladders

读取 Question Ladders。根据 claim 类型选择:

  • Ownership ladder;
  • Architecture ladder;
  • Metric ladder;
  • Benchmark ladder;
  • Failure ladder;
  • Product / user ladder;
  • Open-source ladder;
  • AI-assisted coding ladder;
  • Behavioral ladder。

不要机械把每个 ladder 全问一遍。目标是找到 knowledge / evidence boundary。


4. Live Drill mode

默认交互模式一次只问一个问题。

流程:

Question 1
→ user answer
→ evaluate answer
→ choose next question based on weakness
→ continue 3–7 layers

不要在第一个问题前把完整标准答案全部泄露,否则失去诊断价值。

After each answer

内部判断:

  • factual consistency;
  • technical depth;
  • ownership clarity;
  • evidence link;
  • decision quality;
  • failure awareness;
  • communication。

下一题应该针对最薄弱处,而不是固定脚本。


5. One-shot prep mode

如果用户明确不要 live drill,而是“一次性准备”,则对每个问题给:

Likely question
What the interviewer is testing
Answer skeleton
Evidence to mention
Boundary to state
Dangerous answer
Possible follow-up

仍然不能替用户虚构具体经历。


6. Ownership defense

高风险 title / verb 必须追:

  1. 你本人具体做什么?
  2. 哪些 decision 是你的?
  3. 队友做什么?
  4. AI / coding agent 做什么?
  5. 你如何验收?
  6. 如果你不在,这个结果哪部分不会发生?

AI coding rule

AI 参与不是 red flag 本身。

真正的 red flag:

  • 不知道生成了什么;
  • 不会验证;
  • 不知道失败原因;
  • 无法解释 architecture;
  • 把 AI 输出直接当个人 technical depth。

用户应该能够说明 human judgment:problem definition、spec、review、debug、evaluation、tradeoff。


7. Metric / benchmark defense

所有提升数字至少追:

  • metric 是什么?
  • baseline 是什么?
  • sample size?
  • fixed criteria?
  • repeated trials?
  • variance?
  • model / dependency version?
  • regression?
  • causal claim 是否过强?

Example

Resume:

Improved completion from 61% to 78%.

Follow-up:

这 17 个百分点里,retry、context policy、model variance 各贡献多少?你怎么知道?

如果用户无法回答,可能不是项目没价值,而是这条 bullet 应改成更安全的 measurement language。


8. Architecture defense

要求用户能:

  • 画 data flow;
  • 说明 component boundary;
  • 解释 why;
  • 给 alternative;
  • 说 failure / bottleneck;
  • 说 10× scale 下变化。

“用了 Redis / LangChain / Kafka / vector DB”不是 architecture explanation。


9. Failure defense

至少选一个项目追问:

  • 最先失败什么?
  • 怎么复现?
  • 第一个修复为什么不行?
  • 最后方案带来什么 tradeoff?
  • 现在还剩什么 failure?

真正做过项目的人通常能讲具体失败。

没有 failure narrative 的完美项目要提高怀疑程度。


10. Behavioral questions must remain evidence-backed

行为题也应来自真实经历。

例如:冲突、失败、影响他人、压力、模糊需求。

用:

Situation
Constraint / tension
Decision
Action
Result
Learning / changed behavior

不要把行为题变成“背一套积极故事”。如果结果并不完美,也可以真实表达 decision quality。


11. Scoring

按 Defense Scoring Rubric 评估:

  • Technical Depth;
  • Ownership;
  • Evidence Quality;
  • Decision Quality;
  • Failure Understanding;
  • Communication。

同时给 claim-level verdict:

  • DEFENSIBLE;
  • DEFENSIBLE_WITH_BOUNDARY;
  • KNOWLEDGE_GAP;
  • EVIDENCE_GAP;
  • OWNERSHIP_GAP;
  • DOWNGRADE;
  • REMOVE。

数字评分只有在有足够回答样本时才给,避免假精确。


12. Red flags

明确指出:

  • 会讲名词,讲不出实现;
  • 会讲结果,讲不出 measurement;
  • team result 讲不出个人动作;
  • “主导”没有关键 decision;
  • benchmark 没有 task / metric / baseline;
  • AI 生成代码,用户不会验证;
  • 精确数字没有来源;
  • production claim 实际只有 local demo;
  • 比赛名次描述不准确;
  • 每个回答都像通用 STAR 模板,缺少具体 tension。

13. Corrective action

发现问题后不要教用户“圆”。

选择:

KNOWLEDGE_GAP

  • 回项目 / 源码补理解;
  • 重新跑一次;
  • 画 architecture;
  • 写 failure analysis。

EVIDENCE_GAP

  • /proof 补审计;
  • /build 产生 benchmark / deployment;
  • /contributor 获 external review。

OWNERSHIP_GAP

  • 把 wording 改成实际个人动作;
  • 保留团队结果但明确团队边界。

DOWNGRADE

直接建议 /position 改简历。


14. Default output

Interview Risk Map

ClaimRiskFirst attackEvidence gap

Live Drill

一次一个问题。

Defense Report

claim:
verdict:
technical_depth:
ownership:
evidence_quality:
decision_quality:
failure_understanding:
communication:
strongest_answer:
weakest_answer:
missing_knowledge:
missing_evidence:
recommended_action:
resume_wording_change:

Priority Fixes

只给最重要的 1–3 个问题,按“最可能导致面试失败 × 修复成本”排序。


15. Handoff

按 Cross-skill Handoff Contract:

claims_tested:
  - claim:
    verdict:
    knowledge_gap:
    evidence_gap:
    ownership_gap:
resume_changes:
proof_actions:
build_actions:
recommended_next_skill: /proof | /position | /build | /contributor | /offer

如果 claim 无法支持,优先降低表述或补证据,而不是训练用户更像在说真话。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.