500 AI Agents：Resume Parser 與 Candidate Fit 評分提示詞

Stage 1 System:
Extract structured information from this resume and return JSON:
{
"name": "full name",
"email": "email or null",
"phone": "phone or null",
"location": "city, country or null",
"linkedin": "URL or null",
"github": "URL or null",
"summary": "2-3 sentence professional summary",
"years_experience": number,
"current_title": "current/most recent job title",
"skills": {
"languages": ["Python", "JavaScript", ...],
"frameworks": ["Django", "React", ...],
"tools": ["Docker", "Git", ...],
"soft_skills": ["leadership", ...]
},
"experience": [{"title": "...", "company": "...", "duration": "...", "highlights": ["..."]}],
"education": [{"degree": "...", "institution": "...", "year": "..."}],
"certifications": ["..."],
"languages_spoken": ["English", ...]
}
Return only valid JSON.

Stage 1 User:
{resume_text}

Stage 2 System:
Given this candidate profile and job description, return JSON:
{
"fit_score": 0-100,
"fit_label": "Excellent|Good|Fair|Poor",
"strengths": ["matching point 1", "matching point 2", ...],
"gaps": ["missing skill 1", ...],
"recommendation": "Hire|Consider|Pass",
"recommendation_reason": "2-3 sentence explanation"
}
Return only valid JSON.

Stage 2 User:
Candidate profile:
{candidate_profile_json}

Job description:
{job_desc}