CareerProof
wyl000bdml-sys/CareerProof
Vault de evidências de carreira com matching job-fit e resumes tailorizados
Ver no GitHubSkill de agente + templates para construir um vault de evidências de carreira, combinando projetos reais com descrições de vagas e gerando resumes direcionados sem inventar experiência. Foca em proveniência, confidencialidade e análise truthful de lacunas de competência.
Instalação
npx convez-ai wyl000bdml-sys/CareerProof1 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
- Construir inventário estruturado de projetos com métricas e fontes de evidência
- Analisar job descriptions e mapear habilidades requeridas vs. experiência real
- Gerar variantes de resume direcionadas por track sem overclaiming
- Estratégia de candidatura em múltiplas vagas com gap analysis
Veredito

CareerProof preenche um gap real entre automação de aplicação e tailoring de resume ao focar em proveniência e verdade. A abordagem de vault estruturado com regras anti-overclaim é sólida para profissionais que precisam de credibilidade sem fabricação, apesar de exigir disciplina inicial na catalogação de evidências.
README
A tradução para o português deste README ainda está na fila.
CareerProof
Agent skill + templates for turning real experience into evidence-backed career fit.
CareerProof is a privacy-first agent skill and template system for building a career evidence vault, matching real project experience to job descriptions, and generating truthful resume drafts without inventing experience.
Positioning
CareerProof is not mainly another "click-to-apply" bot. The stronger public angle is:
A provenance-aware career evidence vault that turns resumes, slides, papers, project folders, and job descriptions into truthful, job-specific resume variants and application strategy.
It ships as both a reusable agent skill under skills/careerproof/ and a set of portable YAML/prompt templates.
Use As An Agent Skill
CareerProof can be used directly as a skill by copying:
skills/careerproof/
into an agent skill directory such as:
~/.codex/skills/careerproof/
~/.claude/skills/careerproof/
Then ask the agent:
Use $careerproof to build an evidence-backed career vault and tailor my resume to this job description.
Why this is different
Popular open-source job-search tools usually fall into two buckets:
- Application automation:
- AIHawk / Jobs Applier AI Agent
- LinkedIn Easy Apply bots
- LinkedIn / Indeed automation forks
- Resume matching and tailoring:
- Resume Matcher
- Claude Code resume-tailoring skills
- small LLM resume tailor apps
This template focuses on the missing middle:
- multi-source career extraction, not only one master resume
- evidence-backed project inventory, not free-form rewriting
- job-family strategy, not one JD at a time
- confidentiality-aware metric wording
- truthful gap analysis and anti-overclaim rules
- reusable resume variants by target track
Public repo scope
Recommended public repository name:
career-evidence-vaultjob-fit-evidence-vaulttruthful-resume-tailorcareer-vault-template
Recommended license:
- Use MIT or Apache-2.0 only if the implementation is written cleanly from scratch.
- If reusing code from AGPL projects such as Auto_job_applier_linkedIn or AIHawk-derived code, keep the derivative repository AGPL-compatible.
Core workflow
- Ingest sources:
- existing resumes
- project notes
- slides and papers
- portfolio links
- optional GitHub repositories
- Build a structured career vault:
- projects
- skills
- metrics
- evidence sources
- confidentiality level
- ownership boundaries
- Analyze job descriptions:
- required skills
- preferred skills
- domain signals
- seniority and visa/location constraints
- Match jobs to evidence:
- direct match
- transferable match
- weak or unsupported claim
- missing evidence
- Generate outputs:
- targeted resume outline
- ATS-friendly bullet candidates
- gap analysis
- interview talking points
- optional DOCX resume
What should not be public
Do not publish:
- browser profiles
- cookies or login states
- secrets or API keys
- application logs
- raw private resumes
- private production datasets
- generated resumes with personal contact info
- company-confidential project details
Feature comparison
| Project | Stars checked | Main strength | Gap this template fills |
|---|---|---|---|
| feder-cr/Jobs_Applier_AI_Agent_AIHawk | 29,901 | Large-scale AI job application automation | Weak provenance and deep career-evidence modeling |
| GodsScion/Auto_job_applier_linkedIn | 2,405 | LinkedIn Easy Apply automation and configurable bot | Public version is automation-first, not career-vault-first |
| srbhr/Resume-Matcher | 27,373 | Resume scoring, keyword highlighting, UI, PDF export | Mostly master-resume/JD flow, less multi-source evidence extraction |
| varunr89/resume-tailoring-skill | 512 | Strong truthful tailoring, discovery, confidence scoring | Skill-oriented, less domain-specific evidence schema and privacy packaging |
| beatwad/LinkedIn-AI-Job-Applier-Ultimate | 110 | Multi-platform automation, dashboard, Telegram, anonymization | Automation-heavy; higher breakage and platform-policy risk |
Verdict
The current local work is not more advanced than top GitHub projects in browser automation, UI polish, or community maturity.
It is potentially more advanced in a narrower and more valuable niche:
- PhD/research career evidence extraction
- project-to-job evidence mapping
- truthful technical resume tailoring
- confidentiality-aware metrics
- reusable job-family strategy
That niche is strong enough for a public template repository if the project is refactored into a clean, privacy-safe package.
Proposed public architecture
career-evidence-vault/
README.md
LICENSE
pyproject.toml
examples/
candidate_profile.example.yaml
project_inventory.example.yaml
job_target.example.yaml
career_vault/
schemas/
ingest/
matching/
generation/
privacy/
prompts/
evidence_extraction.md
job_match_scoring.md
truthful_resume_tailoring.md
gap_analysis.md
templates/
resume/
reports/
tests/
Release plan
v0.1: Template-only
- YAML schemas
- prompt templates
- example project inventory
- manual workflow guide
- no automation login
- no real personal data
v0.2: Local CLI
- parse YAML career vault
- score JD against projects
- generate Markdown resume draft
- generate gap analysis report
v0.3: Document generation
- DOCX export
- multiple resume variants
- project evidence matrix
v0.4: Optional integrations
- GitHub project ingestion
- Google Scholar/publication ingestion
- job board scraping in collect-only mode
- no automatic submission by default