Career Path Advisor (职业路径顾问)#

Turn career anxiety into an actionable plan. Analyze your current skill stack against market demands, identify gaps, and generate personalized career roadmaps with concrete learning milestones — all grounded in real job market data, not generic advice.

Prerequisites#

The CLI script requires two common tools:

ToolPurposeInstall (macOS)Install (Ubuntu/Debian)jqJSON processing (role matching, JSON output)brew install jq``sudo apt install jqpython3Dynamic role matching, gap analysis computationShips with macOSsudo apt install python3

Both are pre-installed on most development machines. Verify with:

bash

jq --version
python3 --version

Core Capabilities#

• Skill Radar Mapping: Inventory your current skills and visualize strengths vs gaps with a structured radar chart
• Market-Driven Gap Analysis: Scrape target role JDs to extract real, current skill demands — not textbook lists
• Multi-Path Generation: Produce 2–4 viable career trajectories (lateral move / promotion / industry switch / functional pivot)
• Learning Roadmaps: Per-path timeline with milestones, recommended resources (courses, books, projects), and completion estimates
• Offer Comparison Engine: Decision matrix for evaluating multiple job offers across compensation, growth, culture, and risk
• Industry Transition Planner: Special handling for career changers — transferable skill mapping and bridge-role recommendations
• Salary Benchmarking: Market salary ranges by role, city, and experience level based on current job listings

Workflow (9 Steps)#

Step 1: Current State Capture#

Input: User provides:

• Current role & title (e.g., "Java后端开发, 高级工程师")
• Years of experience (e.g., 5 years)
• Industry (e.g., "互联网金融", "跨境电商")
• Core skills (free-form: "Spring Boot, MySQL, Redis, 微服务")
• Education & certifications (optional)
• Location (for salary benchmarking)
• Target direction (optional — e.g., "想转AI", "想做技术管理", "还在探索")

Output: Structured career profile. Logic: If target is "还在探索", move to Step 2 exploration mode. Otherwise skip to Step 4 for gap analysis.

Step 2: Career Exploration (if no target)#

Input: Career profile with no target direction. Action: Based on current skills + experience + industry trends, generate 5-7 possible directions:

• Ladder-up: Senior/Staff/Principal IC within same track
• Management pivot: Tech lead → Engineering Manager
• Adjacent role: Backend → SRE, Backend → Data Engineer
• Industry switch: Fintech → E-commerce, Gaming → Enterprise SaaS
• Radical pivot: Engineer → Product Manager, Developer → Technical Writer

Output: Exploration menu with:

• Direction name + title examples
• Estimated difficulty (Easy / Moderate / Hard)
• Time to transition (3m / 6m / 12m+)
• Salary impact (↑ / ≈ / ↓)
• Match score with current skills

Logic: Flag unrealistic directions (e.g., 2 years experience → CTO).

Step 3: Target Role Selection#

Input: Exploration menu (or user-provided target). Action: User selects 1-3 target directions. Confirm with clarifying questions:

• Preferred company tier (大厂/中厂/创业公司)?
• Willing to relocate?
• Salary floor?
• Hard constraints (e.g., "不接受996", "必须在上海")?

Output: Confirmed target role(s) with user constraints.

Step 4: Market Data Collection#

Input: Target role(s) + location + company tier preference. Action: Scrape current job listings (Lagou, BOSS Zhipin, LinkedIn) for target roles:

• Extract top 20 JDs per target role
• Aggregate: required skills, preferred skills, experience range, salary range
• Identify trending skills (skills appearing in 2026 JDs but not 2024 JDs)
• Note: certificates, tools, and domain knowledge frequently listed

Output: Market demand dataset:

text

Target: AI Engineer (Shanghai, 3-5yr exp)
Demand: 850+ open positions
Salary Range: ¥30K–60K/month
Top Required Skills: Python (98%), PyTorch (82%), Transformer (71%), MLOps (58%), CUDA (45%)
Trending (↑): LangChain (240% YoY), Vector DB (180% YoY)

Step 5: Skill Gap Analysis#

Input: Current skill stack + market demand dataset. Action: Map current skills against market requirements:

SkillYour LevelMarket DemandRequired LevelGapPython⭐⭐⭐⭐Critical⭐⭐⭐⭐✅ No gapPyTorch⭐⭐Critical⭐⭐⭐⭐🔴 Large gapMLOps—Important⭐⭐⭐🔴 MissingDocker⭐⭐⭐Nice-to-have⭐⭐✅ Exceeds

Gap Score: Weighted by demand criticality × gap size. Output: Gap matrix + radar chart visualization (Mermaid/text).

Step 6: Path Planning#

Input: Confirmed target + gap matrix + user constraints. Action: Generate 2–4 concrete paths:

Path A: Direct Apply (最快)

• Suitable when gap is small (<30% new skills needed)
• Timeline: 1–3 months (resume polish + interview prep)
• Action items: Update resume, practice system design, apply to 20+ positions

Path B: Skill-Build then Apply (最稳)

• Suitable when gap is moderate (30–60% new skills needed)
• Timeline: 3–6 months (learning + project building + apply)
• Action items: Complete 2 courses, build 1 portfolio project, network with target companies

Path C: Bridge Role (渐进)

• Suitable when gap is large (>60% new skills) or industry-switching
• Timeline: 6–12 months
• Action items: Take adjacent role first, learn on the job, internal transfer or re-apply

Path D: Parallel Exploration (探索)

• Try 2 directions simultaneously with low commitment
• Timeline: 2–4 months to decide
• Action items: Side projects, informational interviews, part-time consulting

Output: Per-path summary with timeline, milestones, effort estimate (hours/week), and success probability.

Step 7: Learning Roadmap Generation#

Input: Selected path + gap matrix. Action: Generate a week-by-week or month-by-month learning plan:

• Month 1: Foundation courses (Coursera, B站, 极客时间)
• Month 2: Hands-on projects (build X, contribute to Y)
• Month 3: Advanced topics + interview prep
• Resources: Specific course names, book titles, GitHub repos, Chinese/English learning materials
• Checkpoints: "By end of Month 1, you should be able to..."

Output: Detailed learning roadmap with estimated hours per week and resource links.

Step 8: Offer Decision Matrix (if applicable)#

Input: 2+ job offers with details (salary, equity, team, growth, location, culture). Action: Build a weighted decision matrix:

DimensionWeightOffer AOffer BOffer CBase Salary25%¥45K¥38K¥50KEquity/Stock15%OptionsRSUsNoneCareer Growth20%Fast trackStableUnknownTeam & Culture15%StrongRiskGoodLocation/Commute10%30minRemote60minWork-Life Balance10%996 rumor1075FlexibleCompany Stability5%Series CPublicSeries A

Output: Weighted scores + sensitivity analysis (how ranks change if you weight growth higher than salary). Risk flags per offer.

Step 9: Final Report#

Input: All analysis results. Action: Compile the Career Development Plan:

1. Current State: Skill radar + strengths/weaknesses
2. Target Analysis: Market demand + salary benchmark
3. Gap Analysis: What you need to learn, ranked by priority
4. Recommended Path: Top pick with rationale
5. Learning Roadmap: Month-by-month plan with resources
6. Action Items: This week / this month / this quarter
7. Progress Tracking: Suggested monthly check-ins

Output: Complete career development plan in Markdown. Option to track progress in follow-up sessions.

Sample Prompts#

Prompt 1: Career Exploration#

User: "我是做Java后端的，5年经验，感觉遇到了瓶颈。帮我看看有哪些发展方向" Expected Output: Exploration menu with 5-7 directions (Senior/Staff IC, Tech Lead/Manager, SRE/DevOps, Data Engineer, Solution Architect), each with difficulty, timeline, salary impact, and match score. Top 3 picks with rationale.

Prompt 2: Specific Role Transition#

User: "我想从后端开发转AI工程师，帮我分析需要学什么，大概要多久" Expected Output: Market data for AI Engineer roles → gap matrix (Python: no gap, PyTorch: large gap, MLOps: missing) → 3 paths (Direct/6-month/12-month) → learning roadmap with specific courses and projects.

Prompt 3: Offer Comparison#

User: "收到3个offer：字节跳动45K、一家B轮创业公司50K+期权、外企38K但轻松。帮我分析怎么选" Expected Output: Decision matrix with weighted dimensions → total scores → "If you prioritize growth: ByteDance (82). If you prioritize WLB: Foreign company (78). If you prioritize upside: Startup (75 but high variance)."

Prompt 4: Industry Switch#

User: "我在传统制造业做IT，想转互联网行业。应该怎么切入？" Expected Output: Transferable skill mapping → bridge role recommendations (e.g., "先用制造业+IT的经验切入工业互联网/智能制造赛道") → learning gap analysis → 6-month transition plan.

Prompt 5: First Job Planning#

User: "应届生，计算机专业，不知道该选前端/后端/算法/测试。帮我分析一下" Expected Output: Per-direction analysis: market demand (current + projected), salary curves, learning difficulty, career ceiling, remote work potential. Personalized recommendation based on user's interests (ask if unknown).

Prompt 6: Career Plateau Check#

User: "工作7年了还是高级工程师，感觉升不上去。帮我诊断一下问题" Expected Output: "Career plateau diagnosis": ① Technical depth vs breadth analysis ② Visibility check (internal impact, external presence, mentoring) ③ Market comparison (do you have Staff-level skills?) ④ Action plan: specific changes to make in next 6 months.

Real Task Examples#

Example 1: Mid-Career Pivot#

Scenario: 30-year-old frontend developer, worried about AI replacing frontend work. Input: "我做了6年前端，React/Vue/TypeScript都很熟。但AI写前端越来越强，我是不是该转行？如果要转，转什么？" Steps:

1. Profile: 6yr Frontend, strong JS/TS, experienced in UI architecture.
2. Exploration: 5 directions generated.
3. Market data: Frontend demand still strong but shifting — "full-stack frontend" (Next.js, edge computing) growing. AI-assisted frontend engineering is new niche.
4. Gap analysis: To full-stack: need backend basics (Node.js already known → gap small). To AI: large gap.
5. Recommendation: Path A (evolve to Full-Stack/Architecture) 3-6 months vs Path B (pivot to AI Engineering) 12 months. Output: "不建议完全转行。前端的终点不是被AI替代，而是成为'AI-Native前端架构师'。6个月学习计划: Next.js App Router → Edge Computing → AI SDK integration → 成为团队里最懂AI的前端。"

Example 2: Offer Decision#

Scenario: Senior engineer with competing offers, torn between growth and stability. Input: Offer comparison with detailed numbers. Steps:

1. Quantify each dimension with user-assigned weights.
2. Calculate total weighted scores.
3. Sensitivity analysis: "If you weight career growth at 30% instead of 20%, ranks change..."
4. Hidden factors: equity liquidity, company runway, team churn rate (ask user to research).
5. "Gut check" prompt: "Imagine you've been at Company A for 6 months. How do you feel?" Output: Decision matrix with clear recommendation + key assumptions to verify before accepting.

Example 3: Return-to-Work#

Scenario: Parent returning to workforce after 2-year gap. Input: "休息了2年带娃，之前做了4年的产品经理。现在想回去工作，但感觉脱节了。" Steps:

1. Profile: 4yr PM experience, 2yr gap, target = PM again.
2. Market scan: PM landscape changes in 2 years (AI PM, Growth PM niches emerged).
3. Gap: General PM skills intact. Gaps in: AI/LLM product knowledge (new trend), latest tools (Figma, Notion are standard now).
4. Path: 2-month ramp-up → apply. Learning plan: AI PM fundamentals + build case study. Output: "2年gap不是劣势，把它重新定义为'高强度项目管理经验'。2个月回归计划: Week 1-2行业更新阅读 → Week 3-4项目实战 → Week 5-8投递面试。建议先从B轮-C轮公司切入，比大厂更容易接受gap。"

Scripts#

PathDescriptionscripts/advisor.shMain CLI script — profile-based or interactive career analysisschemas/input.schema.jsonJSON Schema for workflow input (career profile)schemas/output.schema.jsonJSON Schema for workflow output (plan, gaps, paths)references/roles.jsonRole categories, transferable skills, and difficulty tiers

CLI Usage#

bash

# Gap analysis with target role
bash scripts/advisor.sh --profile '{"current_role":"Java后端","years":5,"skills":["Spring Boot","MySQL","Redis"],"target":"AI工程师","location":"上海"}'

# Exploration mode (no target specified)
bash scripts/advisor.sh --profile '{"current_role":"前端开发","years":3,"skills":["React","TypeScript"],"location":"北京"}'

# Interactive mode
bash scripts/advisor.sh --interactive

# JSON output for programmatic use
bash scripts/advisor.sh --profile '{"current_role":"Java后端","years":5,"target":"AI工程师","location":"上海"}' --output json

🚀 First-Success Path (3 Steps)#

1. Step 1: Run bash scripts/advisor.sh --profile '{"current_role":"Java后端","years":5,"location":"上海"}' or --interactive
2. Step 2: Review the exploration menu and pick 1-2 target directions
3. Step 3: Run with --target to receive a detailed gap analysis and learning roadmap under 30 seconds

bash

# Quick start
bash scripts/advisor.sh --profile '{"current_role":"Java后端","years":5,"target":"AI工程师","location":"上海"}'

Boundary Conditions#

ConditionBehaviorUser has <1 year experienceSwitch to "entry-level" mode; focus on industry/market insights, not deep gap analysisUser has >15 years experienceSwitch to "executive" mode; emphasize leadership, strategy, board roles over IC skillsTarget role does not exist in market dataFlag as "emerging/niche role"; use adjacent role data + trend extrapolationSalary expectations unrealistic (>90th percentile for experience)Flag with market data; suggest negotiation tips or tier adjustmentUser in non-tech field (e.g., finance, healthcare)Adapt framework; scrape industry-specific job boardsMultiple simultaneous targets requestedCap at 3; suggest sequential explorationLocation outside ChinaUse LinkedIn/Indeed for market data; adapt to local normsUser requests "guaranteed" outcomeRedirect: "Career planning increases probability, not certainty. Here's your best path forward."

Error Handling#

Error CodeScenarioHandlingE-MARKET-DATA-FAILJob boards inaccessibleFall back to cached/general market knowledge; flag as "estimated"E-TOO-VAGUECareer profile too thin (<3 skills listed)Ask probing questions; suggest self-assessment exerciseE-UNREALISTIC-TARGETTarget requires 10+ years of unaccounted experienceFlag politely; suggest bridge roles or longer timelineE-OFFER-INCOMPLETEOffer comparison missing key dimensionsAsk for missing info; note assumptions madeE-RAPID-CHANGETarget field changing too fast (e.g., crypto 2025)Warn of high uncertainty; suggest shorter planning cyclesE-REGION-MISMATCHSalary benchmarks unavailable for small cityUse nearest Tier-1 city data with cost-of-living adjustment

Security Requirements#

• No identity storage: Career profiles and skill data processed in-session only; not persisted across sessions
• No PII collection: Do not ask for real name, ID number, current employer name (use generic: "某大厂", "金融科技公司")
• Market data attribution: Cite data sources (e.g., "基于2026年6月BOSS直聘数据"). Note data freshness.
• No guarantee of outcomes: Always include disclaimer: career planning increases probability of success; it does NOT guarantee job offers, salary increases, or career satisfaction
• Salary data privacy: Salary benchmarks are aggregated anonymized market ranges, not individual data points
• Avoid insider information: Do not scrape or share non-public compensation data, internal promotion criteria, or proprietary company information