From resume-coach
Guides users through structured coaching to improve resumes for target jobs: recruiter analysis, ideal candidate generation, gap comparison to uncover strengths beyond AI rewriting.
npx claudepluginhub xavierchoi/xavierchoi_skills --plugin resume-coachThis skill uses the workspace's default tool permissions.
Transform resumes from generic AI-polished documents into compelling, personalized stories that pass recruiter screening. This skill implements a coaching methodology that helps users discover and articulate their unique value proposition.
Tailors resumes to job descriptions by analyzing requirements, reordering sections, rephrasing bullets, and adding keywords while preserving all facts truthfully. Use for adapting resumes to target roles.
Tailors resumes to specific job postings by fetching details from URLs, parsing requirements/keywords, mapping candidate experience, and identifying gaps.
Generates tailored resumes for job applications: researches company/role, surfaces undocumented experiences via discovery, matches from resume library, outputs MD/DOCX/PDF while preserving facts.
Share bugs, ideas, or general feedback.
Transform resumes from generic AI-polished documents into compelling, personalized stories that pass recruiter screening. This skill implements a coaching methodology that helps users discover and articulate their unique value proposition.
Most AI-assisted resume improvements produce similar results because they lack context. This skill takes a different approach: instead of simply rewriting, it guides users through a structured discovery process that uncovers hidden strengths and creates genuine differentiation.
Core Philosophy: The goal is not to write the resume FOR the user, but to help them rediscover and reframe their own experiences.
Before starting the process, collect:
If either is missing, request it before proceeding.
Analyze the resume from a hiring manager's viewpoint. Identify 3-5 questions a recruiter would ask when reviewing this resume.
Question Types to Generate:
Implementation:
Use AskUserQuestion to present questions one at a time.
Collect answers to enrich the resume context.
Store responses for use in final resume.
Based on the job posting, generate a fictional "ideal candidate" resume. This represents what the hiring manager imagines as the perfect fit.
Include in the ideal candidate:
Present to User:
## Ideal Candidate Profile
Based on this job posting, here's what the hiring manager's
"dream candidate" might look like:
[Generated ideal candidate resume summary]
This helps us understand what we're competing against.
Compare the user's resume against the ideal candidate. Create a structured comparison that reveals strengths and areas for improvement.
Comparison Format:
| Area | Ideal Candidate | Your Resume | Analysis |
|------|-----------------|-------------|----------|
| Industry Experience | Competitor A, B | Similar Industry C | Transferable |
| Core Skills | X, Y, Z | X, Y | Highlight Z experience |
| Achievements | 50% revenue growth | Project completion | Quantify impact |
Output:
This phase uses the high-spec-generator subagent.
Why a Subagent? The subagent operates without seeing the original resume details, only:
This isolation prevents anchoring to the original resume's framing, producing fresh perspectives on how achievements could be presented.
Invoke the Subagent:
Use your environment's agent invocation method. In Claude Code, use Task tool with subagent_type="resume-coach:high-spec-generator".
IMPORTANT: Do NOT include original resume content in the prompt. Only provide:
This isolation prevents anchoring to the original resume's framing.
Compare the high-spec version against the original resume. Help the user identify expressions and framings they can legitimately adopt.
Present Discoveries:
## Expressions from High-Spec Version
The high-spec version uses these compelling framings:
1. "Drove 30% increase in team efficiency"
→ Do you have any similar efficiency improvements?
2. "Led cross-functional initiative spanning 3 departments"
→ Did you work across teams? How many stakeholders?
3. "Implemented data-driven decision framework"
→ Any analytical approaches you introduced?
Use AskUserQuestion: For each compelling expression, ask if the user has similar experiences they haven't highlighted.
Synthesize all collected information into the final resume:
Quality Criteria:
Output Options:
This skill heavily uses AskUserQuestion for interactive coaching. Follow these patterns:
Fallback: If AskUserQuestion is unavailable in your environment, present one question at a time as a numbered list and wait for user response before proceeding to the next question.
For Recruiter Questions (Phase 1):
For Expression Validation (Phase 5):
Use your environment's task tracking tool to track progress through phases. In Claude Code, prefer TaskCreate/TaskUpdate.
Update task status as each phase completes.
# [Name]
## Contact
[Email] | [Phone] | [LinkedIn] | [Location]
## Summary
[2-3 sentences highlighting key value proposition aligned with job posting]
## Experience
### [Job Title] | [Company]
[Start Date] - [End Date]
- [Achievement with metric from user's answers]
- [Responsibility aligned with job requirements]
- [Project highlighting relevant skills]
### [Previous Position]
...
## Skills
[Skills matching job posting requirements]
## Education
[Relevant education]
This skill uses one subagent:
high-spec-generator: Generates a competitive-level resume version using only company names, titles, and dates. Located in agents/high-spec-generator.md.
Missing resume: "이력서를 먼저 공유해주세요. 파일 경로나 텍스트를 직접 붙여넣어도 됩니다."
Missing job posting: "어떤 포지션에 지원하시나요? 채용공고 URL이나 내용을 공유해주세요."
Incomplete answers: If user skips questions, note the gap and proceed. Missing context will be reflected in final output quality.