convez.ai
SkillProdutividadeIntermediário
sebuzdugan

AIEngineerCV

sebuzdugan/AIEngineerCV

CV builder agnóstico para AI engineers — web, CLI e Claude em uma codebase

Ver no GitHub
InicialPluginScore · de 0 a 100

Construtor de CV especializado para engenheiros de IA com três interfaces: web (drag-and-drop), CLI e skill Claude. Transforma materiais brutos (CV antigo, LinkedIn, GitHub, texto livre) em um CV polido e orientado para o cargo aplicando lógica de contratação em IA, tudo processado localmente sem backend.

Instalação

$npx convez-ai sebuzdugan/AIEngineerCV

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
3
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 4 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Engenheiros de IA criando portfólio diferenciado para seleção técnica
  • Rápida reciclagem de CV em múltiplos formatos (markdown, HTML print-ready)
  • Avaliação de viabilidade de candidato com scoring de recrutador IA

Veredito

Ferramenta focada e bem arquitetada para nicho específico (AI engineers). Promessa de privacy-first (sem backend, dados locais) é cumprida; implementação em TypeScript/Vite é cleancode. Limitação: skill Claude não substitui plugins nativos em 2025.

README

A tradução para o português deste README ainda está na fila.

AIEngineerCV

An open-source CV builder for AI engineers. One brain, three doors.

License: MIT Stars PRs welcome Live

Try it in your browser - no signup, no upload.

Turn your raw materials - an old CV, a LinkedIn export, a GitHub, or just free text - into a polished, role-targeted CV, by applying baked-in AI-engineering hiring judgment instead of generic resume tips.

AIEngineerCV web app

  • One canonical Profile - a validated schema that fully describes a candidate.
  • Three thin adapters - Web (drag-and-drop), CLI (npx/aicv), and a Claude plugin - that each do one job: populate a valid Profile.
  • One generation spec - the deterministic, prompt-driven brain that turns a Profile into a CV.

Bring your own key. Calls go directly from your machine to your model provider. We never proxy, store, or see your key or your data. No backend, no database, no telemetry.

CLI (aicv)

The CLI wraps the shared brain end to end: ingest your materials, answer only the gap questions, generate, score, export.

# from a clone of this repo
pnpm install && pnpm build

export ANTHROPIC_API_KEY=sk-ant-...   # optional: enables the LLM parse + rewrite

aicv init                 # interactive interview, writes profile.json
aicv ingest my-cv.pdf     # parse a CV / LinkedIn export / notes into profile.json
aicv generate             # build ./out/cv.md (LLM rewrite if a key is set, else deterministic)
aicv score                # AI Recruiter Score + guardrail + top-3 fixes
aicv export --md --html   # write ./out/cv.md and a print-ready ./out/cv.html (Print -> PDF)

Every LLM step is optional. With no key, aicv still scores you, renders a clean CV deterministically, and exports it - the score never needs a key. The guardrail (aicv generate) declines out-of-scope CVs unless you pass --force, and never bypasses silently.

Claude plugin

Install it as a one-line Claude Code plugin, then run /cv (or say "build my AI engineer CV"):

/plugin marketplace add sebuzdugan/AIEngineerCV
/plugin install aiengineercv@aiengineercv

The claude-skill/ plugin bundles a /cv slash command and the aiengineer-cv skill, which runs the same flow as the CLI: ingest -> draft Profile -> ask only the gaps -> guardrail -> generate -> score -> write the CV to a file (and a print-ready cv.html; for PDF it shells out to the CLI in Claude Code, or print/web elsewhere). SKILL.md points at the same six expertise assets that power every adapter - the most direct expression of "it works as a spec in Claude."

Web app

The live drag-and-drop app (apps/web) runs the same brain in your browser: drop a CV or paste your background, see the live editable CV preview and a screenshot-worthy AI Recruiter Score, copy the share-card, and export to Markdown or PDF - all client-side. Bring your own key (optional) for the AI parse and rewrite; the score and the deterministic CV need no key. It's a Vite + React static build deployed free to GitHub Pages with zero server cost.

🚧 Status: all four adapters are built - core brain, CLI, Claude plugin, and web app. See docs/superpowers/specs for the design, packages/core for the assets that make the tool good, apps/cli for the CLI, claude-skill for the Claude plugin, and apps/web for the web app.

Built by @sebuzdugan

An AI engineer, for AI engineers. Say hi or follow along:

X YouTube Medium GitHub

MIT licensed. PRs welcome - especially to the taxonomy and rubric; every contribution to the AI-engineering vocabulary makes the tool sharper.

Ver o original no GitHub

Bom para

Relacionados