resume-intelligence-hub

📖 Documentation & landing page → (available in English · 简体中文 · 日本語 · 繁體中文)

An opinionated AI-agent skill that bootstraps a personal career intelligence hub — a private Git repo that is both (a) your single source of truth (resumes, interview prep, grant applications, credential dossiers, pre-submission verification) and (b) your career compass — the seat you want 1–2 levels up, the explicit gap between you and that seat, and what to close this quarter. Works in any AI IDE that reads AGENTS.md (Claude Code, Cursor, Codex, Cline, Windsurf, GitHub Copilot, ...).

Unlike most career skills in the ecosystem — which are one-shot generators ("paste a JD, get a resume") — this one is a framework. It scaffolds a persistent, version-controlled hub and then powers ongoing job search, interview prep, grant applications, and fact-checking on top of it. Output is only half the value — the other half is direction: the hub forces you to name a stretch target, surface the capability gap (skills, title, scope, comp), and revisit it every time you apply, so your applications compound toward a destination instead of drifting.

Who is this for?

• Anyone consolidating scattered resumes/CVs into a single source of truth
• Anyone aiming to jump 1–2 levels up but hasn't yet laid out the capability gap between their current seat and the target seat, or doesn't yet know which target seat to aim for
• Senior professionals (director / VP / CTO / partner / PI) who apply infrequently but each submission is high-stakes
• Bilingual job searchers (Chinese / English hubs both supported, pick one at bootstrap)
• Researchers who also job-hunt (optional dual track: recruitment + grant applications)
• Professionals preparing evidence-heavy credential, licensing, professional-title, or promotion dossiers
• Anyone helping a friend or family member set up the same system

Works for any industry — software, healthcare, finance, law, design, academia, sales, manufacturing, etc. The framework is industry-agnostic; vocabulary customizes via bootstrap Q&A.

What you get

1. A private Git repo with this structure

.
├── profiles/              # Single source of truth
│   ├── master.md          # Job-resume source
│   ├── research.md        # Grant-application source (optional)
│   ├── skills.md          # Competency matrix
│   └── stories.md         # STAR story library
├── jobs/
│   ├── templates/
│   ├── market-watch/      # Target companies + hiring signals
│   └── applications/{company}-{role}-{YYYY-MM-DD}/
├── verification/          # Pre-submission fact-checking
│   ├── credentials/       # Non-sensitive originals
│   ├── references.md      # Path-reference index (sensitive originals stay local)
│   └── {date}-web-check.md
├── assessments/           # Personality tests, 360 feedback
├── credential-applications/ # Optional: title/license/promotion dossiers
├── resumes-archive/
├── research-archive/      # (optional)
├── todo.md / changelog.md
└── AGENTS.md              # AI agent behavior guide

2. Seven workflows the agent executes on the hub

• Career planning & gap analysis — name a stretch target 1-2 levels up, diff the capability gap (skill / scope / credential / network), turn it into a quarterly SMART plan. Run first after bootstrap, re-run every ~quarter.
• JD sourcing — find good JDs (active web search by stretch target, or triage a JD you have)
• JD-tailored resume — generate a custom resume from profiles/master.md against a specific JD
• Interview prep — predict questions, prep STAR answers, tech review, behavioral strategy
• Pre-submission verification — cross-check every load-bearing claim against public sources
• Grant application — generate research proposals in NSFC / NIH / provincial / industry formats (optional)
• Credential / promotion dossier — build an eligibility map, evidence matrix, and safe wording for professional titles, licenses, board certifications, or internal promotion packets (optional)

3. Referenced frameworks

The skill cites established career frameworks by name so users have vocabulary to research further:

• Google's XYZ formula (Laszlo Bock) — resume bullet structure
• STAR — behavioral interview answers
• BEI (McClelland) — behavioral interviewing methodology
• Heilmeier Catechism (DARPA) — research proposal framing
• T-shaped / π-shaped skills — positioning
• Career capital (Cal Newport) — career capital + gap analysis against the target seat
• Stretch target heuristic (1-2 levels up, 1.2-3x comp, 70% match) — application targeting
• Triangulation — due diligence / verification

Install

Via the Skills CLI (recommended):

npx skills add Zenine/resume-intelligence-hub -g -y

Or clone manually into your IDE's skills directory (Claude Code path shown — adjust for Cursor / Codex / Cline / Windsurf):

git clone https://github.com/Zenine/resume-intelligence-hub ~/.claude/skills/resume-intelligence-hub

Use

In your AI IDE, say any of:

• "I've got 5 resumes across different folders — consolidate them into one hub" — bootstrap
• "I want to jump Senior→Staff — diff the gap and give me a plan for this quarter" — career planning & gap analysis
• "A friend sent me this Staff PM JD — triage it, and if it's worth applying, tailor my resume against it" — JD triage + tailored resume
• "Surface stretch-target roles in my space posted in the last two weeks" — active JD sourcing
• "Onsite this Thursday — predict likely questions from this JD and drill me on STAR answers" — interview prep
• "Before I hit submit, cross-check every load-bearing claim in my resume against my GitHub, papers, and LinkedIn" — pre-submission verification
• "Draft an NSFC / NIH grant application profile" — grant application (if research track enabled)
• "Help me prepare a credential or promotion dossier" — credential / promotion dossier (if enabled)

The skill's SKILL.md file instructs the agent on the full bootstrap flow: 8-question interview (language, existing materials, industry, seniority, research track yes/no, credential/promotion track yes/no, resume output language, repo location), then scaffold, then next-steps punch list.

Design philosophy (the non-obvious choices)

1. Single source of truth in profiles/; archives are read-only
2. Positioning lock in AGENTS.md top — change target once, every resume biases to it
3. Multi-track separation when enabled — recruitment, research, and credential dossiers stay distinct
4. Path-references for sensitive originals — repo stays shareable, originals stay local
5. Pre-submission public-source cross-check for high-stakes submissions
6. todo.md / changelog.md split — todo is pending-only, completed items migrate
7. Per-application folders dated — same company twice = two folders
8. STAR stories separate from facts — crafted retelling vs. raw data
9. Monolingual output — pick one of Chinese / English at bootstrap, don't mix
10. AI-IDE agnostic — uses AGENTS.md cross-IDE convention
11. Attribution boundary — personal ownership, team outcomes, and company outcomes are labeled separately; forecasts and pipeline are not treated as revenue

Complementary skills in the ecosystem

This is a framework. For specific one-shot tasks, compose these on top:

Skill	Purpose
paramchoudhary/resumeskills@resume-ats-optimizer	ATS deep optimization
paramchoudhary/resumeskills@linkedin-profile-optimizer	LinkedIn profile
composiohq/awesome-claude-skills@tailored-resume-generator	Alternative JD-tailored generator
anthropics/knowledge-work-plugins@interview-prep	Anthropic official interview prep
refoundai/lenny-skills@career-transitions	Lenny's career transitions
aradotso/trending-skills@awesome-phd-cv	PhD CV (research track)

Privacy

This skill contains no personal data. All templates use placeholder-free prose ("to be filled" markers) that AI fills through conversation with each user. The example STAR stories in templates/{cn,en}/profiles/stories.md are generic fictional scenarios, clearly labeled as anonymized format references.

The hub itself, once populated, contains real personal data — it's designed to live in a private Git repo, never public.

License

MIT. See LICENSE.

Contributing

PRs welcome — especially for:

• Additional language support (currently Chinese + English)
• More industry-specific vocabulary in templates
• New workflows (e.g., salary negotiation, compensation benchmarking, networking playbook)
• Regional JD-sourcing channels (current coverage: China, US, UK, EU, global tech; gaps: SE Asia, India, Japan, Korea, Latam)

Attribution

The frameworks referenced in this skill are the work of their respective authors: Laszlo Bock (Google), David McClelland (BEI), George Heilmeier (DARPA), Cal Newport (So Good They Can't Ignore You). The skill merely cites and applies them — original works should be consulted directly for depth.