/setup
Set up your job search profile. Paste your resume or answer a few questions. Takes 5 minutes. Needed before evaluating jobs.
$ npx -y skills add andrew-shwetzer/career-ops-plugin-do-not-fork-currently-updating-v2- --agent claude-codeHow it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/setup
Context preview
What this command does when you run it.
Set up your job search profile. Paste your resume or answer a few questions. Takes 5 minutes. Needed before evaluating jobs.
Command definition
setup.mdname: setup
description: "Set up your job search profile. Paste your resume or answer a few questions. Takes 5 minutes. Needed before evaluating jobs."
argument-hint: "[--reset to start over]"
user-invocable: true
allowed-tools:
- Read
- Write
- Glob
Set Up Your Profile
Walk the user through creating their job search profile. Conversational, friendly, zero jargon. This is the first thing a new user does.
Step 0: Check Existing
Read `data/profile.yml`. If it exists and no --reset flag:
> "You're already set up! Here's a quick summary of your profile: > > **{Name}** — {Current title} at {Company} > Looking for: {Target role} > > Want to update it? Say 'setup --reset' to start fresh, > or tell me what to change and I'll update just that part."
If --reset flag or file doesn't exist, continue to Step 1.
Step 1: Welcome
> "Let's set up your job search profile. This helps me evaluate jobs, > tailor resumes, and write messages that actually sound like you. > > **Fastest way:** Paste your resume below (text or PDF content) and > I'll pull everything from it automatically. > > **Or** I can ask you a few questions instead. What do you prefer?"
Step 2A: Resume Ingestion (preferred path)
If the user pastes resume content:
1. Parse the resume into structured data:
- Name, contact info, LinkedIn URL
- Work history (each role: title, company, dates, bullet points)
- Education (degree, school, year)
- Skills (extract all mentioned skills and tools)
- Certifications/licenses (if any)
- Projects or portfolio items (if listed)
2. Save the raw resume text to `data/resume.md` as a reference document.
3. Ask follow-up questions for fields NOT in the resume:
- "What kind of role are you looking for next?"
- "Where are you willing to work? (Remote, specific city, flexible)"
- "What's your target salary range? (Skip if you'd rather not say)"
- "Anything else I should know? Career change, gap to explain, special situation?"
4. Continue to Step 3.
Step 2B: Conversational Collection (fallback)
If the user prefers questions, ask these one at a time. Wait for each answer.
1. "What's your name?" 2. "What do you do right now? (Job title and company, or 'between jobs')" 3. "Walk me through your last 2-3 roles briefly. For each: title, company, how long, and one or two things you accomplished." 4. "What kind of role are you looking for? (Job title, industry)" 5. "Where are you willing to work? (City, remote, hybrid, relocate)" 6. "How many years of work experience total?" 7. "What are your strongest skills? (Top 5-10)" 8. "Any certifications or licenses? (Skip if none)" 9. "What's your salary range? (Target and minimum, or skip)" 10. "Got a LinkedIn profile URL?" 11. "Anything else? (Career change, gap, special circumstances, portfolio)"
Step 3: Build Profile
Construct `data/profile.yml` from collected data. Follow the schema in references/profile-schema.md exactly.
Key rules:
- Populate `work_history` with actual role details, not just titles
- Extract quantified achievements into `proof_points`
- Auto-detect persona modifiers:
- If graduated within last 2 years: `recent_graduate: true`
- If previous roles are in a different industry than target: `career_changer: true`
- If gap > 1 year in work history: `career_returner: true`
- If visa_status is anything other than citizen/permanent resident: `international: true`
- Generate `narrative.headline` from their experience (one compelling line)
- Generate `narrative.superpowers` from their strongest skills/achievements
Write the completed profile to `data/profile.yml`.
Step 4: Confirm
Show the user a summary:
> "Here's what I have: > > **{Name}** — {Current title} at {Company} > **Experience:** {years} years > **Looking for:** {Target role} in {industries} > **Location:** {preference} > **Key skills:** {top 5} > **Salary target:** {range} > > **Work history:** > - {Role 1} at {Company} ({dates}) > - {Role 2} at {Company} ({dates}) > - ... > > This look right? I can fix anything now, or you can update later."
Wait for confirmation. Fix any corrections.
Step 5: Next Steps
> "You're all set! Here's what to do next: > > **Option 1:** Paste a job posting (URL or text) and I'll evaluate how well > you match. > > **Option 2:** Say 'scan [company name]' to search their career page for > openings. > > **Option 3:** Say 'help' to see everything I can do."
Create `data/applications.md` if it doesn't exist:
# Job Applications
| Date Added | Date Applied | Company | Role | Score | Status | Evaluation | Notes |
|---|---|---|---|---|---|---|---|
Read more
name: setup description: "Set up your job search profile. Paste your resume or answer a few questions. Takes 5 minutes. Needed before evaluating jobs." argument-hint: "[--reset to start over]" user-invocable: true allowed-tools: - Read - Write - Glob
Set Up Your Profile
Walk the user through creating their job search profile. Conversational, friendly, zero jargon. This is the first thing a new user does.
Step 0: Check Existing
Read `data/profile.yml`. If it exists and no --reset flag:
> "You're already set up! Here's a quick summary of your profile: > > **{Name}** — {Current title} at {Company} > Looking for: {Target role} > > Want to update it? Say 'setup --reset' to start fresh, > or tell me what to change and I'll update just that part."
If --reset flag or file doesn't exist, continue to Step 1.
Step 1: Welcome
> "Let's set up your job search profile. This helps me evaluate jobs, > tailor resumes, and write messages that actually sound like you. > > **Fastest way:** Paste your resume below (text or PDF content) and > I'll pull everything from it automatically. > > **Or** I can ask you a few questions instead. What do you prefer?"
Step 2A: Resume Ingestion (preferred path)
If the user pastes resume content:
1. Parse the resume into structured data:
- Name, contact info, LinkedIn URL
- Work history (each role: title, company, dates, bullet points)
- Education (degree, school, year)
- Skills (extract all mentioned skills and tools)
- Certifications/licenses (if any)
- Projects or portfolio items (if listed)
2. Save the raw resume text to `data/resume.md` as a reference document.
3. Ask follow-up questions for fields NOT in the resume:
- "What kind of role are you looking for next?"
- "Where are you willing to work? (Remote, specific city, flexible)"
- "What's your target salary range? (Skip if you'd rather not say)"
- "Anything else I should know? Career change, gap to explain, special situation?"
4. Continue to Step 3.
Step 2B: Conversational Collection (fallback)
If the user prefers questions, ask these one at a time. Wait for each answer.
1. "What's your name?" 2. "What do you do right now? (Job title and company, or 'between jobs')" 3. "Walk me through your last 2-3 roles briefly. For each: title, company, how long, and one or two things you accomplished." 4. "What kind of role are you looking for? (Job title, industry)" 5. "Where are you willing to work? (City, remote, hybrid, relocate)" 6. "How many years of work experience total?" 7. "What are your strongest skills? (Top 5-10)" 8. "Any certifications or licenses? (Skip if none)" 9. "What's your salary range? (Target and minimum, or skip)" 10. "Got a LinkedIn profile URL?" 11. "Anything else? (Career change, gap, special circumstances, portfolio)"
Step 3: Build Profile
Construct `data/profile.yml` from collected data. Follow the schema in references/profile-schema.md exactly.
Key rules:
- Populate `work_history` with actual role details, not just titles
- Extract quantified achievements into `proof_points`
- Auto-detect persona modifiers:
- If graduated within last 2 years: `recent_graduate: true`
- If previous roles are in a different industry than target: `career_changer: true`
- If gap > 1 year in work history: `career_returner: true`
- If visa_status is anything other than citizen/permanent resident: `international: true`
- Generate `narrative.headline` from their experience (one compelling line)
- Generate `narrative.superpowers` from their strongest skills/achievements
Write the completed profile to `data/profile.yml`.
Step 4: Confirm
Show the user a summary:
> "Here's what I have: > > **{Name}** — {Current title} at {Company} > **Experience:** {years} years > **Looking for:** {Target role} in {industries} > **Location:** {preference} > **Key skills:** {top 5} > **Salary target:** {range} > > **Work history:** > - {Role 1} at {Company} ({dates}) > - {Role 2} at {Company} ({dates}) > - ... > > This look right? I can fix anything now, or you can update later."
Wait for confirmation. Fix any corrections.
Step 5: Next Steps
> "You're all set! Here's what to do next: > > **Option 1:** Paste a job posting (URL or text) and I'll evaluate how well > you match. > > **Option 2:** Say 'scan [company name]' to search their career page for > openings. > > **Option 3:** Say 'help' to see everything I can do."
Create `data/applications.md` if it doesn't exist:
# Job Applications | Date Added | Date Applied | Company | Role | Score | Status | Evaluation | Notes | |---|---|---|---|---|---|---|---|
A Claude Cowork plugin that turns your job search into a system. Evaluate job postings, generate ATS-optimized resumes, scan career portals, track applications, and more.

