Research-backed skills for writing resumes. Every rule traces back to a real study.
Data-driven rules for resume content, drawn from 500k+ hiring outcomes and 10k+ ATS scans. Explains why certain wording hurts (e.g. buzzwords lower pass rate, passive language signals subordination) so the agent can apply the mental model, not just a word list.
Use when:
- Writing or rewriting resume bullets or sections
- Improving word choice, quantification, or phrasing
- Tailoring a resume to a job description (keywords, tech-in-context)
- Structuring or formatting a resume (headers, dates, promotion stack, visual hierarchy)
- Checking for gendered language bias or aspirational filler
Covers: strategic framing, bullet composition (XYZ, So what?, show don't tell), bullet priority and length, quantification and scale anchoring, ATS-tested word choice with causal framing, tailoring and keyword placement, structure and formatting (including de-junioring, interests), visual hierarchy (F-pattern, white space), gendered language
Every time AI rewrites your resume, the layout breaks. Resumx fixes that by keeping your resume in Markdown, where content and layout are separate. AI edits freely, the PDF comes out perfectly fitted every time.
Use when:
- Creating or editing a Resumx resume
- Configuring tags, views, or multi-language output
- Styling, custom CSS, or page fitting
- Using the CLI or converting from other formats
Covers: syntax, CLI, frontmatter, style options, icons, tags, views, variables, multi-language, page fitting, validation, custom CSS
For editors that support agent skills (Claude, Cursor, GitHub Copilot, etc.):
npx skills add resumx/skillsCopy and paste this as the first message in your prompt (before uploading your resume):
Read and follow the resume writing rules from this URL: https://raw.githubusercontent.com/resumx/skills/main/skills/resume-writing/SKILL.md
Rewrite my resume bullets using those rules.
| Study | Finding |
|---|---|
| NBER 30886 | Better writing = 8% more hires, 10% higher wages (n ≈ 500k) |
| Laszlo Bock / Work Rules! | XYZ formula: "Accomplished [X] as measured by [Y] by doing [Z]" |
| TheLadders Eye Tracking (2012, updated 2018) | Recruiters spend 6-7.4 seconds on initial resume scan |
| MatchRate | ATS pass rates from 10k+ scans: action verbs ("Led" 96%), buzzwords ("Responsible for" 23%), passive ("helped" 33%, "involved in" 21%), tech-vague ("coding" 45%, "computer skills" 12%) |
| JobEase / Fortune 500 A/B test | Keywords in first 50 words → 340% more interview requests (n = 8,000) |
| ScoutApply | Job title match → 10.6x interview rate |
| EDLIGO (1,000 rejected resumes) | Vague isolated skill claims → 67% rejection; contextual skills → 34% |
| Inc. / Resume.io | 1,600+ hiring managers flag "proven," "synergy" as filler |
| Nielsen Norman Group | F-pattern reading: eyes scan top horizontally, then down the left margin |
| Forbes / Huntr (59k resumes) | Tailored resumes → 115% higher interview rate vs generic (5.75% vs 2.68%) |
| Resume-Now (18.4M resumes) | Only 10% of resumes include measurable results; 34% of hiring managers call it a dealbreaker |
| ResumeGo (7,712 resumes) | Two-page resumes preferred 2.3x over one-page; scored 21% higher on clarity |
| TalentTuner (944 resumes) | 72% of resumes score below the 70% ATS threshold; average score 57.6% |
| SSRN 4074976 | Gender-incongruent resume language penalizes women's callbacks (330k resumes) |
| ACL 2022 | Gender signals detectable in CVs even after stripping names |
| ICIS 2021 | Neutral verbs outperform both passive and aggressive extremes |