name: career
description: Career analysis - job application assessment with GO/NO-GO scoring, company research, interview prep, and production-ready deliverables. Resume ethics mandatory - never fabricate.
agent: advisor
version: 1.5
classification: public
last_updated: 2026-01-26
effort_default: STANDARD

⛔ DUAL-PATH ROUTING - READ THIS FIRST

STOP. This skill requires the advisor agent for complex requests.

Identity check: If you are NOT the advisor agent AND your request is complex
(opportunity evaluation, GO/NO-GO analysis, application strategy) → DELEGATE NOW:

typescript Task(subagent_type="advisor", prompt="Execute career skill. Request: {user_request}")

DO NOT proceed if you lack advisor expertise for:
- GO/NO-GO opportunity assessment
- Company research and vetting
- Career strategy planning
- Application material development

Path 1 - Simple (Tier 1/Haiku): General career information
- "What's important in evaluating a job offer?"
- "Tell me about compensation negotiation tactics"
- Routes directly, no delegation needed

Path 2 - Complex (Advisor): Opportunity evaluation
- "Should I apply for this role?"
- "Analyze this job opportunity for me"
- Requires advisor delegation

Career Skill

Analyze job opportunities with GO/NO-GO assessment, company research, and tailored application materials.

Supporting Documentation (READ FIRST)

⚠️ MANDATORY: Read these supporting documents BEFORE executing this skill:

1. docs/professional-tone-requirements.md - Formatting standards for all deliverables (resumes, cover letters)
2. docs/skill-tier-integration.md - Model tier routing and cost optimization

These documents contain critical requirements that MUST be followed during execution. Failure to read them will result in deliverables that violate formatting standards or use incorrect model tiers.

Model Tier Routing

This skill uses the three-tier orchestration engine for cost-optimal execution:

Routing Decision Tree:
- Career advice and guidance: Tier 1 (Free) - $0.00 (e.g., "What's the market for X role?")
- Job posting analysis with GO/NO-GO scoring: Tier 2a (Grok 3 Mini) - $0.80/1M ← Primary
- Complex analysis with assumption testing: Tier 2a (Grok 3 Mini) - $0.80/1M (role fit validation)

Why Tier 2a for career analysis?
- Assumption challenging (role fit, career growth)
- Logic validation (GO/NO-GO scoring)
- Gap detection in job descriptions
- Cost savings vs. Sonnet: 67% ($0.80 vs $2.40 per 1M)

How it works:
1. You invoke /career with a job posting or career question
2. Base Claude's tier selector analyzes the request
3. Guidance queries route to Tier 1 (free)
4. Job analysis routes to Tier 2a for assumption testing
5. GO/NO-GO score challenges assumptions before recommending

Cost Model:
- Career guidance: $0 (Tier 1)
- Job posting analysis: ~$0.80 (Tier 2a for assumptions)
- Interview prep: ~$0.40 (Tier 1/2a hybrid)

Reference: READ docs/skill-tier-integration.md for complete skill mapping and tier selection criteria.

Pre-flight Checklist (MANDATORY)

STOP! Before executing this skill:

• [ ] Read this SKILL.md completely
• [ ] Understand the 5-phase workflow
• [ ] Phase 1 (Assess) is the CRITICAL GATE - NO-GO = stop entirely
• [ ] Resume ethics: NEVER fabricate experience, skills, or achievements
• [ ] Professional tone requirements: No AI cliches, emojis, or icons
• [ ] Resume chronology: Present to past (reverse chronological)

Core Philosophy

GO/NO-GO First: Don't waste time on poor-fit roles. Phase 1 determines if you proceed.

Evidence-Based: All claims backed by research with cited sources.

Ethics Mandatory: Optimization = highlighting relevant experience. Fabrication = changing what you did. Never cross that line.

Job Discovery Mode (NEW)

Automatically discover and score jobs from hiring.cafe using a local scraper.

Instead of manually finding job postings, you can now search and score multiple opportunities:

bun skills/career/scripts/discover-jobs.ts "cybersecurity director remote" --days 14 --min-score 60

Features:
- Scrapes hiring.cafe with your search query
- Scores jobs against your resume using GO/NO-GO methodology
- Returns top matches ranked by fit percentage
- No API keys or authentication required

How it works:
1. Scraper finds jobs matching your query
2. Extracts job details (title, company, description, URL)
3. Scores each job against your resume (0-100%)
4. Displays ranked results with recommendations

See: skills/career/scripts/scrapers/README.md for full documentation.

Critical Rules

1. GO/NO-GO is the Gate - Score <60% = STOP. Don't proceed with poor-fit roles.
2. Resume Ethics - NEVER fabricate experience, skills, certifications, or achievements
3. Cite Sources - All research cited inline per section with URLs
4. No Guessing - If gaps exist in resume, ASK USER. Never assume undocumented experience.
5. Professional Tone - No AI cliches, emojis, or icons. Complete sentences only.
6. Resume Chronology - Work experience MUST be reverse chronological (present to past)

Professional Tone Requirements (MANDATORY)

All deliverables must follow professional formatting standards. READ docs/professional-tone-requirements.md for complete requirements BEFORE generating any deliverables.

Key Rules:
- NO emojis, icons, AI cliches, or exclamation marks
- Complete sentences, formal business tone, active voice
- Resumes: Reverse chronological, 2 pages max, ready to send
- Cover letters: 4 paragraphs max, 350 words, ready to send

Resource Auto-Detection

Resume Search Strategy:
1. Check for any resume file in skills/career/input/ directory (.md, .pdf, .docx)
2. If multiple resumes found, ask user which to use
3. If no resume found, ask user to provide path or content

Supported Formats: Markdown (.md), PDF (.pdf), Word (.docx)

5-Phase Workflow

┌─────────┐     ┌──────────┐     ┌─────────┐     ┌──────────┐     ┌─────────┐
│ ASSESS  │────▶│ RESEARCH │────▶│ PREPARE │────▶│ GENERATE │────▶│ DELIVER │
│ (Gate)  │     │          │     │         │     │          │     │         │
└────┬────┘     └────┬─────┘     └────┬────┘     └────┬─────┘     └────┬────┘
     │               │                │               │               │
     ▼               ▼                ▼               ▼               ▼
  GO/NO-GO      10+ Sources      5 Priority      5 Files         Strategy
  Decision      Collected        Areas + Q&A     Created         Provided

CRITICAL: If Phase 1 results in NO-GO (<60%), STOP. Do not proceed to Phase 2.

Phase 1: ASSESS (GO/NO-GO Gate)

The most important phase. Determines if this opportunity is worth pursuing.

Scoring Methodology

Decision Gates

Red Flags (Auto NO-GO)

• Required certification not held (and can't obtain quickly)
• Experience 5+ years below requirement
• Technology stack 0% overlap
• Location requires unwanted relocation

Phase File: phases/01-assess.md

Phase 2: RESEARCH (Company Intelligence)

Gather comprehensive intelligence on the target company.

Research Sources

Research Categories

1. Company Basics - Size, HQ, industry, funding, recent news
2. Leadership - CISO/CIO/CTO backgrounds, priorities
3. Security/Compliance - Breach history, certifications, posture
4. Culture & Sentiment - Glassdoor, employee reviews
5. Tech Stack - Languages, frameworks, cloud providers
6. Recent Developments - Funding, acquisitions, hiring trends

Gate Criteria

• [ ] 10+ sources collected with URLs
• [ ] All 6 research categories covered
• [ ] Sources cited inline per section

Phase File: phases/02-research.md

Phase 3: PREPARE (Interview Preparation)

Develop priority areas and interview Q&A based on research.

Priority Areas (5)

For each priority area:
- Why It Matters - Evidence from research
- Their Need - What they're seeking
- Your Position - How to leverage your experience
- Talking Points - 2-3 key points to emphasize

Interview Q&A (10-12 Questions)

Questions to Ask (5-7)

Strategic questions that demonstrate research and genuine interest.

Phase File: phases/03-prepare.md

Phase 4: GENERATE (Create Deliverables)

Create EXACTLY 6 focused deliverable files.

CRITICAL: All deliverables MUST follow Professional Tone Requirements (see section above).

Deliverables

All filenames lowercase-hyphen following Unix standards. NO CAPS, NO SPACES, NO NUMBERS.

Requirements: Each file standalone and focused. NO 900+ line executive summaries, NO multiple versions, NO ALL CAPS/numbered/spaced filenames, NO extra files beyond the 6 specified.

job-posting.md Template

# [Company] - [Position Title]

## ⚠️ Application Information (CRITICAL)

**Application URL:** [Full clickable URL - TEST THIS LINK]
**Job ID/Requisition:** [ID if available - helps find posting if URL breaks]
**Application Deadline:** [Date if specified]

**Source:** [Indeed, LinkedIn, company careers page, referral]
**Date Found:** [YYYY-MM-DD]
**Posting Date:** [YYYY-MM-DD if available]
**Location:** [City, State]

---

## Job Description

[Full original job posting text - complete, unedited]

---

## Application Requirements

- [Requirement 1: e.g., resume, cover letter]
- [Requirement 2: e.g., portfolio, references]
- [Any specific application instructions]

---

## Additional Notes

[Any relevant notes about the application process, contacts, referrals, etc.]

CRITICAL REQUIREMENTS:
- Application URL must be complete and functional (test the link!)
- Include Job ID/requisition number if available (backup way to find posting)
- Original job text must be complete (postings often get removed later)
- This file is THE source of truth for applying - without it, user cannot apply

Resume Ethics (MANDATORY)

ALLOWED:
- Reorder bullets to emphasize relevant experience
- Reorder positions to ensure reverse chronological order (most recent first)
- Rewrite Professional Summary for this role
- Remove irrelevant positions
- Add keywords from job posting (if you actually have the skill)

FORBIDDEN:
- Change job titles
- Fabricate experience or achievements
- Add technologies you didn't actually use
- Guess at undocumented experience

If gaps exist: STOP and ASK USER before proceeding.

Formatting Enforcement

Before finalizing deliverables, verify:
- [ ] EXACTLY 6 .md files with lowercase-hyphen names
- [ ] job-posting.md has working Application URL (test the link!)
- [ ] No files exceed size limits (match: 200, company: 400, interview: 500 lines)
- [ ] No "executive summary" file that duplicates everything
- [ ] Resume work experience is reverse chronological - VERIFY: First position has most recent end date, last position has oldest end date
- [ ] Resume and cover letter have current date (YYYY-MM-DD format)
- [ ] Resume uses bullet points with proper spacing - no squished dense paragraphs
- [ ] No emojis, icons, or decorative symbols in any file
- [ ] No AI cliches or buzzwords
- [ ] All bullet points are complete sentences or detailed phrases
- [ ] Professional business tone throughout
- [ ] Cover letter is under 350 words with 4 paragraphs
- [ ] Cover letter ends at signature line with NO extra content below
- [ ] Resume has NO extra notes or instructions below main content
- [ ] Both documents are ready to print/send as-is

Output Validation Checklist

Before marking Phase 4 complete, validate:

# Validation requirements
required_files = [
    "job-posting.md",      # Must have Application URL with http/https
    "match-assessment.md", # Max 200 lines
    "company-research.md", # Max 400 lines
    "interview-prep.md",   # Max 500 lines
    "resume.md",           # 1-2 pages
    "cover-letter.md"      # <350 words
]

# Exactly 6 .md files (no extras, no PDFs unless requested)
# All filenames match regex: ^[a-z-]+\.md$
# job-posting.md contains "Application URL:" and "http"
# No file named "executive-summary.md" or "1-EXECUTIVE-SUMMARY.md"

Phase File: phases/04-generate.md

Phase 5: DELIVER (Submission Strategy)

Provide actionable next steps for the user.

Submission Timing

• Best days: Tuesday-Thursday
• Best time: 7-9 AM in company timezone
• Avoid: Friday afternoon, weekends, holidays

Follow-Up Timeline

• Day 3-5: LinkedIn connection to hiring manager
• Week 2: Follow-up email if no response
• Week 4: Final check-in, then move on

Networking Strategy

• Identify 2nd-degree connections
• Informational interview opportunities
• Internal referral possibilities

Phase File: phases/05-deliver.md

Deliverable Finalization Standards

CRITICAL: All output files must be production-ready.

Output Requirements

• Lowercase-hyphen filenames only: resume.md not RESUME.md
• Production-ready: Copy/paste to Word/PDF with no editing needed
• Current dates, clean formatting, no markdown artifacts or AI comments

Output Structure

EXACTLY 6 files. No more, no less.

Naming Convention: All lowercase-hyphen following Unix standards.

skills/career/output/{company}-{role}-{date}/
├── job-posting.md           # Original job description and APPLICATION URL (critical!)
├── match-assessment.md      # GO/NO-GO analysis (100-200 lines, concise)
├── company-research.md      # Company intelligence (200-400 lines)
├── interview-prep.md        # Priority areas and Q&A (300-500 lines - the bulk)
├── resume.md                # Resume optimized for role (1-2 pages)
└── cover-letter.md          # Cover letter (<350 words)

File Naming Rules: All lowercase-hyphen (Unix standards). NO caps, spaces, numbers, versions, or mixed case.

DO NOT create: README/executive-summary files, phases/ subdirectory, multiple versions, PDFs (unless requested), or anything beyond the 6 files listed above.

Content Distribution (IMPORTANT)

DO NOT create a massive "executive summary" that duplicates everything. Content should be distributed as follows:

• job-posting.md: Job description + application URL (THE critical file for applying)
• match-assessment.md: GO/NO-GO decision ONLY (concise, 2 pages max)
• company-research.md: Company intelligence with sources
• interview-prep.md: Priority areas + Q&A + questions (this gets the bulk of interview content)
• resume.md: Production-ready resume
• cover-letter.md: Production-ready cover letter

Each file is standalone, focused, and manageable - no 900+ line monsters.

Progress Tracking

Use TodoWrite to display progress through phases:

[ ] Phase 1: Assess - Calculate match score, GO/NO-GO decision
[ ] Phase 2: Research - Gather company intelligence (10+ sources)
[ ] Phase 3: Prepare - Develop priority areas and Q&A
[ ] Phase 4: Generate - Create 5 deliverable files
[ ] Phase 5: Deliver - Provide submission strategy

Each phase updates status as it progresses:
- pending → in_progress → completed

Model Selection

Default: Sonnet (analysis, document generation)
Research: WebSearch + Grok via OpenRouter (social intel)

Error Recovery

Related Skills

• /clifton - CliftonStrengths analysis (separate, not integrated)
• /mentorship - Learning roadmaps and skill development

Version: 1.5
Last Updated: 2026-01-16
Status: Active