job-search

name	job-search
description	Find relevant roles, compare postings, and identify the best jobs to apply to by matching JD requirements against a resume.

Job Search Skill

Purpose

Use this skill when the user wants help with job discovery, shortlist creation, JD comparison, or deciding which roles are the best fit to apply to based on a resume and multiple JDs.

Outcomes

• Search for roles by title, location, seniority, and domain
• Filter weak-fit postings before deeper analysis
• Normalize multiple JDs into a comparable structure
• Rank target jobs by application fit
• Highlight required skills, preferred skills, experience, and education constraints
• Produce clean text or Markdown ready for JD Matcher

Recommended Workflow

1. Confirm target role, market, location, and must-have constraints.
2. Collect candidate postings from job boards, company career pages, or recruiter messages.
3. Remove duplicated or low-signal listings.
4. Normalize each JD into a compact structure.
5. Compare roles and rank them by expected application fit.
6. Keep the highest-fit opportunities for application.
7. Export the final JD block for use in this repo.

Output Format

# Application Target Shortlist

## Role 1
- Title: Senior Backend Engineer
- Company: Example Inc.
- Location: Remote
- Seniority: Senior
- Experience: 5+ years
- Education: Bachelor or above

### Required Skills
- Go
- Distributed systems
- PostgreSQL

### Preferred Skills
- LLM
- RAG
- Kubernetes

### Clean JD
<normalized JD text>

## Role 2
- Title: AI Engineer
- Company: Sample Labs
- Location: Singapore / Remote
- Seniority: Mid-Senior
- Experience: 3+ years
- Education: Flexible

### Required Skills
- Python
- Prompt engineering
- API integration

### Preferred Skills
- LangChain
- Vector database
- Evaluation

### Clean JD
<normalized JD text>

## Recommendation
- Best target to apply: Role 1
- Why: Stronger overlap on must-have backend skills, lower experience gap, and better alignment with current project history

Guardrails

• Preserve the original hiring intent.
• Do not add qualifications that are not present in the source.
• Mark uncertainty when a posting is vague or incomplete.
• Prefer concise, comparable summaries over long raw copy.

Notes For This Repo

• This skill is documentation-only for now.
• Output should stay compatible with the parser heuristics in src/service/parser.js.
• Use this skill before matching when the user starts from job discovery instead of a ready JD.
• The main problem this skill solves is deciding which target岗位最值得优先投递。