yanliudesign/offer-toolkit-skill / offer-compare-skill
offer-compare-skill
VerifiedOffer 对比决策器。用户同时拿到两份(或多份)offer 纠结怎么选时,扮演 Senior Career Decision Advisor:结构化对比 comp / 成长 / AI 敞口 / 公司强度 / 团队风险 / 晋升速度 / 生活方式,识别 front-loaded vs long-term upside、resume value、switch-out 难度、hidden risks,最后给一条**明确、有立场、不中性**的推荐(并附「你是 X 类人选 A,你是 Y 类人选 B」的分叉判定)。三步:贴两份 offer → 补 priorities & 现状 → 自动生成并打开一份 HTML Offer Decision Report。是 offer-toolkit 里 JD Skill / Resume Skill / BQ Skill 拿到结果之后的下一环。关键词:offer compare, 选 offer, 两个 offer, offer decision, TC compare, 该去哪家, which offer, 4-year TC, RSU, sign-on, front-loaded, long-term upside, career decision。
| File | Kind | Size |
|---|---|---|
| LICENSE | other | 1.0 KB |
| README.md | markdown | 9.2 KB |
| README.zh.md | markdown | 9.1 KB |
| SKILL.md | skill md | 10.7 KB |
| examples/offer-compare-template.html | other | 60.8 KB |
| frameworks/dimension-rubric.md | markdown | 5.8 KB |
| frameworks/hidden-signals-dictionary.md | markdown | 6.1 KB |
| frameworks/offer-compare-report.md | markdown | 15.8 KB |
| frameworks/risk-taxonomy.md | markdown | 6.0 KB |
| offer-bank/_index.md | markdown | 3.0 KB |
| offer-bank/_template.md | markdown | 2.3 KB |
| offer-bank/anthropic-vs-linear-staff-pd-202607.md | markdown | 3.9 KB |
| offer-bank/microsoft-stay-vs-linear-staff-pd-202607.md | markdown | 5.6 KB |
| offer-bank/msft-vs-linear-vs-anthropic-pd-202607.md | markdown | 5.7 KB |
| prompts/dimension-compare.md | markdown | 4.8 KB |
| prompts/hidden-signals.md | markdown | 4.1 KB |
| prompts/offer-breakdown.md | markdown | 5.2 KB |
| prompts/recommendation.md | markdown | 5.0 KB |
| prompts/risk-analysis.md | markdown | 3.8 KB |
prompts/dimension-compare.md
View on GitHub# Prompt · Dimension Compare
**目的:** 在 7 个维度上分别给 Offer A / Offer B 一个 0-10 打分 + 一句话解释 + 谁赢。这是报告的 §2。
---
## 输入
- 两份 offer 的 breakdown(来自 [offer-breakdown.md](offer-breakdown.md))
- 用户的 priorities(如果有)
- 用户当前处境(visa / burnout / 家庭)
- 公司背景(能查就查,查不到标 `[需用户补充]`)
---
## 7 个维度(按顺序)
参考评分标准:[../frameworks/dimension-rubric.md](../frameworks/dimension-rubric.md)
### 1. 💰 Compensation Quality
**不只看 4-year TC 总额**,还看:
- 每年现金流健康度(base + bonus)
- Equity 组成风险(refresh 有无 / 股价假设)
- Sign-on clawback 期
- 相对当前 base 的涨幅
**输出模板:**
```
A: 8.5/10 — $1.2M mid-4yr TC · cash-heavy · low equity risk
B: 7.0/10 — $1.4M mid-4yr TC · 但 60% back-loaded 到 Y3/Y4 · 股价假设激进
🏆 A(Winner)— B 名义 TC 更高,但风险调整后 A 更实在
```
### 2. 📈 Career Growth
看:
- Role scope(IC / manager / staff+)
- Level 是否升级(当前 level → offer level)
- 团队规模 · 产品阶段(0→1 / 1→10 / 10→100)
- 未来 12-24 mo 能不能学到新东西 / 拓展新能力
### 3. 🤖 AI / Future Exposure
2026 年是 offer 决策里**最容易被低估**的一维。看:
- 是否在 AI native 团队 / 产品
- 公司整体 AI 战略强度
- 你的 role 会不会被 AI 反噬(如"运营 role"vs"AI product 设计")
- 3 年后这份履历在 job market 上的可迁移性
**这一维必须给分,不许标 N/A。** 就算两家都不做 AI,也要判断"谁离 AI 更近"。
### 4. 🏢 Company Strength
看:
- 财务健康(营收 / 现金流 / 估值趋势)
- 产品市场地位(垄断 / 挑战者 / 长尾)
- 品牌在你 target industry 的认可度
- 24 mo 内被裁员 / 被收购 / IPO 的概率
### 5. 👤 Manager / Team Risk(可推断)
用户很少直接说,但**必须推断给分**(低置信度也要标):
- 团队近期变动(HM 是新上任还是老臣?)
- 面试过程里 HM 的行为信号(决策快 / 追问深 / 尊重你的经历 = 加分;反复变面试轮次 / 招聘周期长 = 减分)
- Glassdoor 上该团队的评论(如能查)
- 公司整体文化健康度(review process 是不是 Amazon 式互撕?是不是 Netflix 式高绩效文化?)
如果完全没信号 → 给 5.0/10(neutral)+ 标 `[信号不足]`。
### 6. 🔁 Promotion Speed
看:
- 公司整体 promo 节奏(Google/Meta 从 senior 到 staff 通常 3-5 年 · Amazon 从 L6 到 L7 更快 · startup 完全看老板)
- 该 level 上下一级 headcount 是否有空位
- Team 是不是新组建(新团队更容易早期升)
- 该 level 是否已经在 "career level"(不需要 promo 就能稳定的 level,如 Google L5 / Meta E5 / Amazon L6)
### 7. 🌍 Location / Lifestyle
看:
- Location cost of living 调整后的净收入
- Commute / on-site 要求(hybrid vs full remote vs 5-day RTO)
- 与家庭 / 配偶职业的兼容
- Visa(H1B transfer / GC sponsorship / 是否 layoff 立刻走人的风险)
- 用户是否 burnout(当前公司 → 新公司 workload 增加?减少?)
---
## 输出格式(塞进报告 §2 的 JSON-ish 结构)
```yaml
dimensions:
- name: "💰 Compensation Quality"
a_score: 8.5
a_note: "cash-heavy · low equity risk"
b_score: 7.0
b_note: "60% back-loaded to Y3/Y4"
winner: "A"
winner_note: "B 名义 TC 更高,但风险调整后 A 更实在"
confidence: "high" # high / medium / low
- name: "📈 Career Growth"
a_score: 7.0
...
```
**Confidence 标注规则:**
- `high` = 有明确数据支持(TC 数字、level 定义、公开公司数据)
- `medium` = 有部分间接信号(Glassdoor 评论、行业口碑)
- `low` = 主要靠推断(如 manager risk 没有面试信号)
---
## 加权总分(可选,报告顶部 gauge 用)
如果用户给了 priorities,按 priorities 加权:
- 用户 top 1 priority 那一维 → ×3
- 用户 top 2 → ×2
- 用户 top 3 → ×1.5
- 其余 → ×1
否则用默认权重(Balanced 模型):
- Comp × 2 · Growth × 2 · AI × 1.5 · Company × 1.5 · Manager × 1 · Promo × 1 · Lifestyle × 1
计算:`A_total = sum(a_score × weight) / sum(weight)` → 0-10 → ×10 = 0-100
**Gauge 两条(A vs B),差距 >8 分 = 明确赢家 · 差距 3-8 = 有倾向 · 差距 <3 = 势均力敌(此时 §5 Recommendation 靠 tie-breaker 决定)。**
---
## 不许做的事
- ❌ 所有维度都 8/10(假中立)— 必须有真实差异
- ❌ 不给 winner(Tie 只在 <0.5 分差时允许)
- ❌ Manager Risk 标 N/A — 必须推断,就算低置信度
- ❌ AI 维度对非 AI 公司标 N/A — 必须判断"离 AI 多远"
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| 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: 472k | Updated | Claude CodeCodex | Unverified | Install |
| 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: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| 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 available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |