Skip to content

yanliudesign/offer-toolkit-skill / job-description-skill

job-description-skill

Verified

Job Description 解码器 + Offer 策略系统。把任何 JD 翻译成「这家公司到底在招什么人 / 我的匹配度 / Gap 在哪 / 为什么投/不投 / 面试会问什么 / 下一步怎么走」。只有 3 步:贴 JD → 给简历 → 自动生成并打开一份 HTML Offer Strategy Report。是整个求职链路的入口,向下衔接 Resume Skill 和 BQ Skill。关键词:job description, JD, 职位描述, 匹配度, ATS, resume tailor, 简历定制, 招聘经理, 投不投, should I apply, 面试预测, offer strategy。

<div align="center">

[中文](./README.zh.md) · **English**

# 🔍 Job Description Skill

---

**Decode any JD → get a one-page Offer Strategy report.** Should I apply · how well I match · top interview questions.

[![License](https://img.shields.io/badge/LICENSE-MIT-4c8bf5?style=flat-square&labelColor=333)](../LICENSE)
[![Version](https://img.shields.io/badge/VERSION-1.0.0-2ea44f?style=flat-square&labelColor=333)]()
[![Sections](https://img.shields.io/badge/SECTIONS-10-2ea44f?style=flat-square&labelColor=333)]()
[![Stars](https://img.shields.io/github/stars/yanliudesign/job-description-skill?style=flat-square&label=STARS&color=e37f2c&labelColor=333)](https://github.com/yanliudesign/job-description-skill/stargazers)

[![Claude Code](https://img.shields.io/badge/Claude_Code-Skill-d97757?style=flat-square&labelColor=1a1a1a&logo=anthropic&logoColor=white)](https://claude.ai/code)
[![Codex](https://img.shields.io/badge/Codex-Skill-2ea44f?style=flat-square&labelColor=1a1a1a)]()
[![OpenCode](https://img.shields.io/badge/OpenCode-Skill-4c8bf5?style=flat-square&labelColor=1a1a1a)]()
[![OpenClaw](https://img.shields.io/badge/OpenClaw-Skill-8b5cf6?style=flat-square&labelColor=1a1a1a)]()
[![Hermes](https://img.shields.io/badge/Hermes-Skill-e879a8?style=flat-square&labelColor=1a1a1a)]()

</div>

> 📦 Part of the **[offer-toolkit-skill](https://github.com/yanliudesign/offer-toolkit-skill)** — the full job-hunt bundle (Search · JD · Resume · BQ · Compare · Negotiate). Install the bundle to get all six in one shot.

A Claude skill for reading job descriptions. Give it a JD and your resume, and it hands back an HTML report telling you: whether this job is worth applying to, how well you match it, where the gaps are, what interviews will probably ask, whether the salary is reasonable, and what to do over the next six weeks. A JD is rarely just a recruiting blurb — it's usually the hiring manager's real expectations wrapped in HR boilerplate, and this skill unwraps that.

---

## How it works — just 3 steps

Invoke the skill with anything like "help me with this job" / "走一下这个 JD" and you go through exactly:

1. **Paste the JD** — link or full text.
2. **Give it your resume** — three options:
   - 🅰️ Upload / paste your full resume (recommended)
   - 🅱️ Paste your full LinkedIn
   - 🅲️ No resume yet → short self-introduction. Results will be flagged as "unverified against a real resume" and you'll be nudged toward **[Resume Skill](https://github.com/yanliudesign/resume-builder-skill)** to build one.
3. **HTML report auto-opens** — a single-file Offer Strategy Report is generated to `~/Desktop/Claude skills/offer-strategy-<slug>.html` and pops open in your browser. Inside: TL;DR verdict, two gauges, the 10 sections below, and in-page Export PDF / Export Markdown buttons.

That's the whole thing. No multi-step wizard, no separate "只解码" / "只算匹配度" outputs to wade through — the 5 internal flows (Decode / Match / Tailor / Predict / Should I Apply) all run silently and assemble into the same HTML report.

If you explicitly want only one flow ("only decode this" / "just match score") the skill skips the report and gives you the raw output for that one flow.

---

## Offer Strategy Report · the 10-section framework

Every wizard run ends with a single-file HTML report at `~/Desktop/Claude skills/offer-strategy-<slug>.html`. Fixed 10-section spine, plus a TL;DR preamble and a key-metrics gauge block:

![Offer Strategy Report preview](docs/report-preview-en-v2.png)

*Hero of an actual run — verdict + assumptions + TL;DR + two gauges, plus the in-page Export PDF / Export Markdown buttons in the bottom right.*

<sub>More screenshots: [English gallery](docs/en/) · [中文相册](docs/zh/)</sub>

### ▶ Video walkthrough (~2 min)

<a href="docs/en/demo.mp4" title="Click to play the full 2-minute walkthrough">
  <img src="docs/en/preview.gif" alt="▶ Play 2-min walkthrough — Anthropic × Product Designer, Claude Code" width="100%">
</a>

<sub>⬆ First 8 seconds auto-play as a preview. <a href="docs/en/demo.mp4"><strong>Click to watch the full 2-minute walkthrough</strong></a> — live run against a real Anthropic JD (paste JD → attach resume → HTML report opens in your browser, ~15 MB, opens in a new tab). 中文版本 · <a href="README.zh.md#▶-视频演示-2-分钟">Chinese walkthrough</a>.</sub>

| # | Section | What it answers |
|---|---------|-----------------|
| — | **TL;DR · one-minute verdict** | Should I apply? ⭐ 1–5. |
| — | **Key metrics** | Match Score gauge + Interview Probability gauge. |
| **1** | **Company background** | Who runs this place, what are the cultural signals, what shipped in the last 12–24 months. |
| **2** | **Product analysis for this role** | What product / surface this role owns + the 4 core design challenges + the *real* scope behind the JD. |
| **3** | **JD deep decode** | Why this JD reads the way it does (JD 解码精华 + Hidden Signal dictionary hits). |
| **4** | **JD ↔ resume line-by-line match** | Must Have cards + full Match Matrix (3 columns: JD ask / resume evidence / score+gap) + Nice + 10-dim Hidden Signal radar. |
| **5** | **Gap — what to upgrade** | 2–4 gap cards with severity (high/middle) + 4-week mitigation. |
| **6** | **Why apply / why not** | 3 reasons each side. Every "why not" comes with an HM probe response. |
| **7** | **Salary reality check** | JD band + market data + negotiation talking points. |
| **8** | **Next step** | Career arc fit + referral playbook (who, DM template) + portfolio audit (4-week action list). |
| **9** | **Top 10 interview questions** | 面试预测预览. Full Top 20 + behavior stories → hand off to BQ Skill. |
| **10** | **6-week action plan** | Week-by-week to-dos: before applying / after applying / before interview. |

The report supports **Export to PDF** (CJK-safe Noto font embedding) and **Export to Markdown** (single-file MD download). Spec: [`frameworks/offer-strategy-report.md`](frameworks/offer-strategy-report.md) · skeleton: [`examples/offer-strategy-template.html`](examples/offer-strategy-template.html).

---

## Three rules

1. **Decode before you act** — Decoding the JD first is mandatory before doing match or tailor. Matching against the wrong target means the match is wrong.
2. **Never fabricate** — Never invent experience, skills, or numbers the candidate didn't claim. If real evidence is missing, mark it ❌.
3. **One thing at a time** — No mixing. If the decode step left ambiguity, clarify before moving on.

---

## File Structure

```
job-description-skill/
├── SKILL.md                      # Entry · routing · three rules
├── prompts/                      # Execution prompts for each flow
│   ├── jd-decoder.md
│   ├── match-score.md
│   ├── resume-tailor.md
│   ├── interview-predictor.md
│   └── should-i-apply.md
├── frameworks/                   # Reusable frameworks / rubrics
│   ├── decode-patterns.md        # JD keyword → real intent dictionary
│   ├── match-rubric.md           # Scoring rubric (Must / Nice / Hidden)
│   ├── resume-tailoring.md       # Three-version resume strategy
│   ├── go-no-go.md               # ⭐ rating + interview probability formula
│   └── offer-strategy-report.md  # Final HTML report — sections, visual system, content rules
├── examples/                     # Reference skeletons (no personal data)
│   └── offer-strategy-template.html
└── jd-bank/                      # Local cache of analyzed JDs (gitignored)
    ├── _index.md                 # Index
    └── _jd-template.md           # Starter template for new JDs
```

---

## How it thinks

- **A JD reflects three layers** — company signals, the hiring manager's real expectations, and the product / role in question. The skill's job is to see through the HR packaging to those three.
- **Decide before you optimize** — most candidates keep polishing resumes for the wrong roles. Decide whether to apply at all first.
- **Every score has a formula** — not magic. Every Match Score and Interview Probability comes from an explicit formula in `frameworks/`. Disagree with the premise? You can argue back.

---

## Related skills

Continue the workflow with these related skills:

- [offer-toolkit-skill](https://github.com/yanliudesign/offer-toolkit-skill) — the all-in-one bundle (Search · JD · Resume · BQ · Compare · Negotiate)
- [job-hunt-skill](https://github.com/yanliudesign/job-hunt-skill) — Job discovery and searchable opportunity list
- [resume-builder-skill](https://github.com/yanliudesign/resume-builder-skill) — Resume Builder & Beautifier (12 print-ready templates)
- [Behavior-question-skill](https://github.com/yanliudesign/Behavior-question-skill) — Behavioral interview / story bank

```
See a dream job → JD Skill (decode · match · should-I-apply)
                     ↓ decide to apply
                  Resume Skill (tailor + polish)
                     ↓ get the interview
                  BQ Skill (mine stories · mock interview)
```

---

## License

MIT — fork it, remix it, ship your own version.

Created by [Dreameryanyan](https://www.linkedin.com/in/yanliudesign/) ·
[LinkedIn](https://www.linkedin.com/in/yanliudesign/) ·
[X](https://x.com/yanliudreamer) ·
[Xiaohongshu](https://www.xiaohongshu.com/notification)
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall