Job Search Skills
kbroadbent/job-search-skills
Sistema end-to-end de busca de emprego com memória compartilhada
Ver no GitHubSistema integrado de skills para Claude que automatiza todo o ciclo de busca por emprego e preparação para entrevistas. Organiza candidaturas, pesquisas de empresas, planos de estudo personalizados e histórias profissionais em um repositório estruturado que persiste conhecimento entre sessões.
Instalação
npx convez-ai kbroadbent/job-search-skills1 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
Casos de uso
- Rastrear múltiplas candidaturas em paralelo com análise de aderência
- Gerar e executar planos de estudo customizados por posição
- Preparar entrevistas técnicas e comportamentais com contexto de empresa e histórias profissionais
- Pesquisar empresas e encontrar contatos de recrutadores para outreach
Veredito

Solução madura e bem-estruturada para profissionais em busca ativa de emprego. Integra pesquisa, documentação, preparação e rastreamento em um modelo coesivo que realmente aproveita memória persistente — raro em skills.
README
A tradução para o português deste README ainda está na fila.
job-search
End-to-end job search system for Claude. A connected set of skills that share a common directory structure and memory model so the agent never starts from scratch — every session loads what it has already learned about you, what's in your pipeline, and what you've studied.
Skills
The skills fall into four groups: setup skills you run once to stand things up, main skills that drive the day-to-day job search, support skills the main flows orchestrate (but that you can also run directly), and a maintenance skill for keeping your inbox in order.
Setup
Run these to stand up the system and plan your prep.
| Skill | What it does |
|---|---|
job-search-init | One-time setup. Walks you through a short interview, then writes the directory scaffolding (config.yaml, CLAUDE.md, profile.md, insights.md, Story Bank.md, study-plans/) every other skill reads from. |
create-study-plan | Generate a customized study plan via interview. Writes weekly plan files into study-plans/{slug}/ and updates the active-study pointer. Supports modify-in-place mode for active plans. |
Main
The skills you reach for as you apply and practice.
| Skill | What it does |
|---|---|
apply-to-job | Track a new position. Writes posting.md and position-fit.md (gap analysis with strengths, gaps, relevant stories, red flags, plus a status header). On Apply intent, also invokes resume-tailor and offers find-recruiter. |
study-session | Run a daily prep session against the active study. Updates progress and appends durable insights to insights.md. |
interview-prep | Prep for a specific upcoming scheduled interview. Pulls calendar context, researches interviewers and reported questions, suggests stories and topics to refresh. |
Support
The main flows call these automatically, but each runs standalone too.
| Skill | What it does |
|---|---|
company-research | Research a company (and optionally a specific posting). Writes companies/{co}/company.md. |
resume-tailor | Tailor your resume to a specific posting. Writes the tailored .docx and .pdf into the position folder. |
find-recruiter | Find a recruiter at a company and draft outreach. Updates companies/{co}/contacts.md and writes the position's outreach.md. |
Maintenance
| Skill | What it does |
|---|---|
label-job-search-emails | Sort recent job-search emails into per-company Gmail sublabels and archive them out of the inbox. |
Directory shape
After running job-search-init, the directory looks like:
{root}/
config.yaml
CLAUDE.md
profile.md
insights.md
Story Bank.md
{your-resume}.docx
companies/
study-plans/
index.md
Positions in flight live under companies/{slug}/{pos}/ — there's no separate pipeline file. Status, recruiter, and posting all live inside the position folder.
As you use the plugin, content accumulates:
companies/
acme/
company.md
contacts.md
staff-backend/
posting.md
position-fit.md
Acme-Staff-Backend-resume.docx
Acme-Staff-Backend-resume.pdf
outreach.md
prep/
recruiter-screen.md
tech-screen.md
system-design.md
leadership.md
interviews/
2026-05-02-tech-screen.md
2026-05-02-tech-screen-debrief.md
study-plans/
staff-backend-4wk/
plan.md
week-1.md ... week-4.md
progress.md
Memory model
Four files are always loaded at the start of every session:
profile.md— who you are, target role, comm styleinsights.md— running collection of patterns, gotchas, and weak spots learned across all studies and interviewsCLAUDE.md— lean session log + pointersstudy-plans/index.md— roster of all study plans with one-line summaries
config.yaml is also loaded for configuration (calendar provider, mail settings, LinkedIn search method, active study). Positions in flight aren't loaded eagerly — they're scanned from companies/ on demand.
This means: when you switch between study plans, start a new search, or come back after a break, the agent already knows you. It doesn't re-learn what's in those files; it picks up where you left off.
Routing: stable identity → profile.md. Recurring patterns and gotchas → insights.md. This-session activity → study-plans/{slug}/progress.md. New stories that surface during practice → Story Bank.md (with help refining them).
Getting started
Install the plugin and run job-search-init to scaffold a new directory. After init, run create-study-plan if you want to start a structured prep plan, or apply-to-job if you have a posting you want to track.