Resumes Are Dead — Your .claude/ Folder Is Your New Career Card

AI blogger Nav Toor made a provocative argument: in 2026, your .claude/ directory — CLAUDE.md, skills, MCP configs — is replacing the traditional resume as a developer’s most important professional artifact. The data backing this claim is hard to dismiss.

The Data: Three Shifts Happening Simultaneously

1. AI Is Writing the Code

2. Junior Jobs Are Disappearing

Junior developer job postings dropped 67% from 2024 to 2025. Many remaining “junior” listings actually require senior-level experience. The entry-level funnel is collapsing — not because juniors aren’t needed, but because the definition of “entry-level work” has changed.

3. Interviews Are Changing

Interviewers no longer test if you can write code from scratch. They observe how you collaborate with AI.

Tyler Folkman (hiring manager) described the new evaluation:

“The key distinction: is this developer casually playing with AI, or using AI to deliver production software? Can they write specs first? Critically review AI output? Write tests? Or do they treat AI as a wish-granting pool — generate and ship without checking?”

The interview questions that matter now:

Your .claude/ Folder as Portfolio

Nav Toor’s argument: your .claude/ directory reveals more about your engineering judgment than any resume line item.

.claude/
├── CLAUDE.md ··············· Your engineering philosophy
│                             - What coding standards do you enforce?
│                             - How do you handle errors?
│                             - What architecture decisions matter to you?
│
├── rules/ ·················· Your quality standards
│   ├── code-style.md         - Do you have opinions about code?
│   └── testing.md            - Do you write tests or skip them?
│
├── commands/ ··············· Your workflows
│   ├── review.md             - How do you review code?
│   └── deploy.md             - What's your deployment process?
│
├── skills/ ················· Your domain knowledge
│   └── domain/SKILL.md       - What expertise have you encoded?
│
└── agents/ ················· Your delegation strategy
    ├── reviewer.md           - How do you break work into roles?
    └── security.md           - Do you think about security?

What Each File Reveals

“I Can Write Python” Means Nothing Now

The punchline from 默庵’s analysis:

“‘I can write Python’ on a resume has lost all differentiating value. What interviewers actually want to see is: can you harness AI to build real, working software?”

This means:

• Writing specs before generating code
• Critically reviewing AI output (not blindly accepting)
• Writing tests (not treating AI as a wish-granting pool)
• Understanding architecture (not just syntax)
• Knowing when AI is wrong (not agreeing with everything it says)

What This Means for Students

This is directly relevant for CS education:

What to Teach

The New Student Portfolio

Instead of a GitHub profile with toy projects, the new portfolio might look like:

1. CLAUDE.md files — showing engineering judgment across different projects
2. Custom skills — domain knowledge encoded as reusable AI skills
3. Automated workflows — commands and hooks showing process maturity
4. Verified AI output — projects with tests proving the student can evaluate AI work

How LearnAI Team Could Use This

• Career readiness modules — Have students build and explain a project-specific CLAUDE.md as evidence of engineering judgment.
• Portfolio reviews — Evaluate whether students can document AI collaboration workflows, verification habits, and domain constraints.
• AI coding workshops — Use the .claude/ folder as a concrete artifact for teaching spec-writing, review discipline, and safe delegation.
• Employer conversations — Translate hiring-market changes into curriculum: students need to show how they direct and verify AI-generated work.

Real-World Use Cases

• Junior developer portfolio — A candidate shares repos with CLAUDE.md, tests, and custom commands showing collaboration with coding agents.
• Mock technical interview — Students walk through how they spec features, review AI output, and prove correctness with tests.
• Capstone project setup — Teams define shared AI rules, commands, and review workflows before implementation begins.
• Career coaching — Advisors help students turn AI-assisted projects into evidence of process maturity and engineering judgment.

The Counterargument

To be fair: Nav Toor’s framing is deliberately provocative. Resumes aren’t literally dead — hiring processes are slow to change, and most companies still use traditional screening. But the direction is clear: the ability to collaborate with AI is becoming more important than the ability to write code alone. Whether that shows up as a .claude/ folder or some other artifact, the signal employers want is changing.

Links

• Nav Toor on X: heynavtoor
• Tyler Folkman on hiring: Referenced in Nav Toor’s analysis
• Related: Junior Dev Resume in the Age of AI (DEV Community)