薪资市场扫描仪
- 作者仓库星标 2,010
- 作者仓库 tinyfish-cookbook
- 领域
- 数据
- 作者声明的智能体
- 未找到明确声明;不据此推断已兼容或已测试
- 静态检查
- 88 / 100 · 启发式扫描,不代表运行安全
- 运行环境证据
- 未执行;当前没有隔离运行测试证据
- 作者 / 版本 / 许可证
- @tinyfish-io · 未声明许可证
- 上下文用量评级
- 低消耗
- 接入复杂程度
- 需简单配置
- 是否需要外部接口密钥
- 需要 · Vendor-specific
- 兼容的系统
- macOS · Linux · Windows
- 运行环境要求
- 未检出明确要求
- 检测到的文件与系统行为
-
- 只读
- 网络行为
- 允许外网请求
- 安装命令数
- 无(仅作为资料)
需要注意: 未声明工具权限范围;实际权限由运行环境决定。
第 1 步 - 收集输入
您需要:职位/角色 - 例如`Senior Software Engineer`、`ML Engineer`、`Product Designer`、`DevOps Engineer` 位置 — 例如`London`、`Singapore`、`San Francisco`、`Remote` 经验年数(可选)——例如`3-5 years`、`senior`、`entry level` 特定公司(可选)— 如果用户想知道特定公司支付的费用
第 2 步 — 并行薪资扫描
同时解雇所有特工。来源因地点而异——包括最相关的来源。在运行之前,替换: `{ROLE}` — 人类可读的,例如`Senior Software Engineer` `{ROLE_ENCODED}` — URL 编码,例如`Senior%20Software%20Engineer` `{ROLE_SLUG}` — Levels.fyi slug 例如`software-engineer`
第 3 步 — 综合市场情况
合并所有来源的数据并计算聚合范围。
# Salary Market Scanner
Scrape live job boards and salary databases to find real compensation data for any role and location — not outdated surveys, but what companies are actually posting and paying right now.
## Pre-flight Check (REQUIRED)
Before making any TinyFish call, always run BOTH checks:
**1. CLI installed?**
```bash
which tinyfish && tinyfish --version || echo "TINYFISH_CLI_NOT_INSTALLED"
```
If not installed, stop and tell the user:
> Install the TinyFish CLI: `npm install -g @tiny-fish/cli`
**2. Authenticated?**
```bash
tinyfish auth status
```
If not authenticated, stop and tell the user:
> You need a TinyFish API key. Get one at: https://agent.tinyfish.ai/api-keys
>
> Then authenticate:
> ```
> tinyfish auth login
> ```
Do NOT proceed until both checks pass.
---
## Step 1 — Gather inputs
You need:
- **Job title / role** — e.g. `Senior Software Engineer`, `ML Engineer`, `Product Designer`, `DevOps Engineer`
- **Location** — e.g. `London`, `Singapore`, `San Francisco`, `Remote`
- **Years of experience** (optional) — e.g. `3-5 years`, `senior`, `entry level`
- **Specific company** (optional) — if the user wants to know what a specific company pays
If location is not provided, ask before proceeding. Salary data varies dramatically by market.
---
## Step 2 — Parallel salary scan
Fire all agents simultaneously. Sources vary by location — include the most relevant ones.
```bash
# Agent 1 — Levels.fyi (best for tech roles, especially US/global big tech)
tinyfish agent run \
--url "https://www.levels.fyi/t/{ROLE_SLUG}/?country={COUNTRY}" \
"You are on Levels.fyi showing compensation data for {ROLE} in {LOCATION}.
Extract:
- Median total compensation
- Base salary range (p25 to p75)
- Bonus range
- Stock/equity range (if shown)
- Sample size (number of data points)
- Top companies listed and their compensation ranges
- Any breakdown by years of experience if visible
STRICT RULES:
- Do NOT click any company or individual entry
- Read only the aggregate data visible on the page
- If no data for this location, return {found: false, reason: 'no data for location'}
Return JSON: {found: bool, median_total, base_p25, base_p75, bonus_range, equity_range, sample_size, top_companies: [{company, base, total}], yoe_breakdown: []}" \
--sync > /tmp/sal_levels.json &
# Agent 2 — Glassdoor salaries
tinyfish agent run \
--url "https://www.google.com/search?q=glassdoor+{ROLE_ENCODED}+salary+{LOCATION_ENCODED}+site:glassdoor.com/Salaries" \
"You are on Google search results. Find the most relevant Glassdoor salary page for {ROLE} in {LOCATION} and click it.
On the Glassdoor salary page extract:
- Median base salary
- Salary range (low to high)
- Number of salary reports
- Additional pay (bonus, profit sharing) range if shown
- Top companies paying for this role if listed
STRICT RULES:
- Click only the first Glassdoor salary result
- Do NOT click any other links after landing on Glassdoor
- Read only the aggregate salary data visible on the page
- If the page asks you to sign in, extract whatever is visible before the gate
Return JSON: {median_base, salary_low, salary_high, report_count, additional_pay_range, top_companies: [{company, salary}]}" \
--sync > /tmp/sal_glassdoor.json &
# Agent 3 — LinkedIn Jobs (extract posted salary ranges from active listings)
tinyfish agent run \
--url "https://www.linkedin.com/jobs/search/?keywords={ROLE_ENCODED}&location={LOCATION_ENCODED}&f_SB2=1&sortBy=DD" \
"You are on LinkedIn job search results for {ROLE} in {LOCATION}, filtered to show salary information, sorted by date.
For each job listing card visible on the page:
- Click into the listing to open the job detail panel on the right
- Look for the salary range in the detail panel (often shown near the top under the job title)
- Extract: job title, company name, salary range, employment type
- Go back to the listing and repeat for the next one
STRICT RULES:
- Only extract listings that show an explicit salary — skip those without
- Maximum 10 listings then stop
- Do NOT navigate away from the search results page
Return JSON array: [{title, company, salary_range, employment_type}]" \
--sync > /tmp/sal_linkedin.json &
# Agent 4 — Indeed salaries
tinyfish agent run \
--url "https://www.indeed.com/career/{ROLE_INDEED}/salaries?from=top_sb&l={LOCATION_ENCODED}" \
"You are on Indeed's salary page for {ROLE} in {LOCATION}.
Extract:
- Average base salary
- Salary range (low to high)
- Number of salary reports
- Salary by experience level (if shown: entry, mid, senior)
- Top paying companies for this role (if listed)
STRICT RULES:
- Do NOT click any links
- Read only the aggregate data on this page
Return JSON: {average_salary, salary_low, salary_high, report_count, by_experience: [{level, salary}], top_companies: [{company, salary}]}" \
--sync > /tmp/sal_indeed.json &
wait
echo "=== LEVELS ===" && cat /tmp/sal_levels.json
echo "=== GLASSDOOR ===" && cat /tmp/sal_glassdoor.json
echo "=== LINKEDIN ===" && cat /tmp/sal_linkedin.json
echo "=== INDEED ===" && cat /tmp/sal_indeed.json
```
**Before running**, replace:
- `{ROLE}` — human-readable e.g. `Senior Software Engineer`
- `{ROLE_ENCODED}` — URL-encoded e.g. `Senior%20Software%20Engineer`
- `{ROLE_SLUG}` — Levels.fyi slug e.g. `software-engineer`
- `{ROLE_INDEED}` — Indeed format e.g. `software-engineer`
- `{LOCATION}` — e.g. `London`, `Singapore`
- `{LOCATION_ENCODED}` — URL-encoded e.g. `London%2C%20England`
- `{COUNTRY}` — country code for Levels.fyi e.g. `GB`, `SG`, `US`
---
## Step 3 — Synthesize the market picture
Combine data from all sources and calculate aggregate ranges.
```
## Salary Market Report — {ROLE} · {LOCATION}
*Live data scraped from Levels.fyi, Glassdoor, LinkedIn, and Indeed*
*{date} · Based on {N} total data points*
---
### 💰 Compensation Summary
| | Low | Median | High |
|---|---|---|---|
| **Base Salary** | {low} | {median} | {high} |
| **Total Comp** (incl. bonus/equity) | {low} | {median} | {high} |
> All figures in {CURRENCY}. "Total comp" includes base + annual bonus + annualized equity where data is available.
---
### 📊 By Experience Level
| Level | Typical Base |
|---|---|
| Entry (0-2 yrs) | {range} |
| Mid (3-5 yrs) | {range} |
| Senior (6+ yrs) | {range} |
| Staff / Principal | {range} |
*(Skip levels where no data was found)*
---
### 🏢 What Companies Are Posting
From active LinkedIn job listings with disclosed salaries:
| Company | Role | Posted Range |
|---|---|---|
| {company} | {title} | {range} |
---
### 🏆 Top Paying Companies
*(From Levels.fyi and Glassdoor)*
| Company | Median Base | Median Total |
|---|---|---|
| {company} | {base} | {total} |
---
### 📈 Market Signals
{2-3 sentences on what the data says about this market — is it competitive, is there a wide spread, are companies being transparent about pay?}
---
### 🔍 Data Sources
- Levels.fyi: {sample_size} data points / not found
- Glassdoor: {report_count} salary reports / not found
- LinkedIn: {N} active listings with disclosed salaries
- Indeed: {report_count} salary reports / not found
```
---
## Edge Cases
- **Levels.fyi has no data for this location** — lean on Glassdoor and Indeed; note that Levels.fyi skews toward US big tech
- **Role title is unusual** — try common variations (e.g. "ML Engineer" → "Machine Learning Engineer", "AI Engineer")
- **Location is a small city** — broaden to the country or nearest major city and note the change
- **Remote role** — scrape for both the user's country and the US market, present both (remote jobs often use US pay bands)
- **Non-tech role** — skip Levels.fyi (tech-only), rely on Glassdoor and Indeed
- **Salary shown in different currencies** — normalize to the local currency and note conversion rate used
- **User is asking if their salary is competitive** — after presenting the data, ask what they're currently earning and give a direct assessment
章节与文件索引
按说明文件中的顺序排列。
飞行前检查(必需)
第 1 步 - 收集输入
第 2 步 — 并行薪资扫描
第 3 步 — 综合市场情况
边缘情况
文件/命令 -> npm install -g @tiny-fish/cli · Senior Software Engineer · ML Engineer · Product Designer · DevOps Engineer · London · Singapore · San Francisco
内容校验值 -> aa175fbe88ab
当前原文的自动摘译,尚未人工审校。
飞行前检查(必需)
在进行任何 TinyFish 调用之前,请始终运行两项检查: 1. CLI 已安装?如果未安装,请停止并告诉用户:安装 TinyFish CLI:
npm install -g @tiny-fish/cli2. 已通过身份验证?边缘情况
Levels.fyi 没有该位置的数据 - 依赖 Glassdoor 和 Indeed;请注意,Levels.fyi 偏向于美国大型科技公司 角色标题很不寻常 - 尝试常见的变体(例如“ML 工程师”→“机器学习工程师”、“AI 工程师”) 位置是一个小城市 - 扩大到国家或最近的主要城市并注意变化 远程角色 - 刮擦用户所在的国家和美国市场,同时提供两者(远程工作通常使用美国薪资带)