Cv Review

CV Review

Quick Start

/cv-review path/to/cv.tex
/cv-review path/to/resume.pdf

Natural language ("review my CV") → ask for file path.

Workflow

1. Read CV

• .tex / .txt / .md → read directly, strip markup
• .pdf → run pdftotext file.pdf - via Bash; if unavailable, use Read tool
• Warn once for PDFs: formatting analysis limited

2. Detect experience level

• Student/new grad: < 1 year total work experience
• Experienced: 1+ years
• Calibrates section order expectations and flagging threshold

3. Detect niche from stack, job titles, project descriptions

• Clear signal → proceed
• Two niches compete → ask: "Your CV shows signals of both [X] and [Y]. Which role are you targeting?"
• Supported: SWE, ML/AI Engineering, Data Science, Systems, Security, DevOps/SRE, NLP/Research
• See [rubric-niches.md](rubric-niches.md)

4. Run base rubric — applies to all niches

• See [rubric-base.md](rubric-base.md)

5. Run niche rubric

• See [rubric-niches.md](rubric-niches.md)

6. Run ATS + structure checks

• See [ats-guidelines.md](ats-guidelines.md)

7. Output (format below)

8. Offer rewrites — after output, ask: "Want rewrite suggestions for any flagged bullets?"

Output Format

NICHE: [niche] | LEVEL: [student / experienced]

SECTION VERDICTS
----------------
[Section Name]    Pass / Weak / Fix — one-line reason if not Pass

TOP FIXES
---------
1. [Specific, actionable fix]
2. ...
(max 5 — only issues a recruiter would notice)

Flagging Rules

• Only flag if recruiter would actually stop or skip because of it
• Calibrate to experience level — thin experience on student CV is not a flaw
• Don't flag absent elements not expected for the role or level
• 3 real fixes beat 8 nitpicks

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: VohraAK
• Source: VohraAK/agent-skills
• License: MIT

Install and usage instructions live in the source repository linked above.