resume-tailor

resume-tailor

A Claude Code skill that converts a job description into a scoring rubric and restructures each resume experience into PHER format.

---

The JD is not a wish list — it's a scoring rubric

Look closely at job postings from IT companies. Next to "We're looking for someone who..." you'll often find:

"Please write your experience in the format: Problem definition → Hypothesis → Execution → Result"

That's not a writing tip. That's the scoring rubric.

resume-tailor extracts the hidden evaluation criteria from a JD and restructures each resume experience to match exactly what the hiring team is grading.

---

Before / After

Before — typical task-listing format

담당: 결제 시스템 개발 및 고도화
- 다양한 PG사 연동 및 결제 플로우 구현

After — PHER structure

[P] Checkout abandonment reached 15% as payment response time exceeded 3 seconds.
    APM data confirmed synchronous PG API calls as the bottleneck.

[H] Hypothesized that converting to async queue-based calls would bring response
    time under 1 second. Chose Kafka over Redis Streams for replay and audit log
    requirements.

[E] Led design and implementation of Kafka-based async payment pipeline with a
    team of 3 backend engineers over 2 months. Personally designed retry logic and
    idempotency key strategy; drove code reviews.

[R] Response time: 3s → 0.8s. Abandonment rate: 15% → 4%.
    Gained hands-on experience with Bulkhead and Circuit Breaker patterns in production.

---

Installation

In the Claude Code terminal:

/plugin install https://github.com/SeokRae/resume-tailor

Or copy the skill directly:

cp -r skills/resume-tailor ~/.claude/skills/resume-tailor

---

Usage

After starting the skill, provide the JD file and resume file paths.

analyze my resume

tailor my resume to this JD @jd.md @resume.md

Output files saved to _workspace/resume-tailor/YYYY-MM-DD-NNN/:

File	Contents
03_gap_report.md	Gap analysis report (🔴/🟡/🟢 classification)
04_interview_notes.md	Deep interview transcript
05_resume_revised.md	Revised resume
05_change_log.md	Full Before/After comparison

---

5-Phase Process

Each phase stops for user confirmation before proceeding.

Phase 1  JD Analysis
         Extract scoring rubric — explicit criteria, hidden signals, keyword map
         Resume writing guide in JD → treated as primary scoring criterion
         ↓
Phase 2  Resume Audit
         Measure PHER completeness of each experience
         JD criteria ↔ experience mapping matrix
         ↓
Phase 3  Gap Analysis
         🔴 Critical (not covered) / 🟡 Weak (evidence lacking)
         🟢 Strong (meets criteria) / ⚪ Irrelevant (no JD value)
         Rewrite priority order
         ↓
Phase 4  Deep Interview
         Gap-type-specific questions to surface details missing from the resume
         Extracts facts from the user — AI never fabricates
         ↓
Phase 5  Rewrite
         Restructure experiences into PHER + Before/After comparison
         6-point validation (fact preservation · JD coverage · PHER · keywords · length · naturalness)

---

PHER Framework

Element	Core question
P Problem	Why did you do this? What problem existed?
H Hypothesis	Why this approach? Why not the alternatives?
E Execution	Who with, what role, which decisions did you own?
R Result	What's the quantified outcome? What did you learn?

---

Absolute Rules

Rule	Description
Preserve facts	Never generate metrics, experiences, or skills absent from the original resume or deep interview
JD criteria first	Optimize for this specific JD — not a generic "good resume" standard
Phase gates	No auto-advancement — user confirms each phase result before proceeding

---

한국어 설명

JD에 숨어있는 평가 기준을 추출하고 이력서 경험을 PHER 구조로 재구성하는 Claude Code 스킬.

핵심 원칙: JD는 채점 기준표다 — JD가 묻는 것에만 최적화된 이력서를 만든다.

사용법: 이력서 분석해줘 또는 JD 기반으로 이력서 고쳐줘 @jd.md @resume.md

출력: 갭 분석 리포트 + 딥 인터뷰 기록 + 개선 이력서 + Before/After 비교

5단계: JD 분석 → 이력서 감사 → 갭 분석 → 딥 인터뷰 → 리라이트 (각 단계 후 사용자 확인)

---

Contributing

PHER examples, interview questions, and JD analysis patterns are welcome. Please read CONTRIBUTING.md first.

• 🐛 Bug report
• ✨ Feature request
• 📝 Add PHER example

---

License

MIT