Career Application Advisor is an open-source agent skill for evidence-based job applications. It helps job seekers review and tailor resumes, identify missing evidence, analyze job descriptions, research employers and roles, compare offers, and reason about career development.
Current release: v1.0.0
- Reviews resumes for positioning, evidence quality, chronology, credibility, and readability.
- Generates prioritized questions before rewriting incomplete experience claims.
- Decodes job descriptions into mission, must-haves, preferences, risks, and interview unknowns.
- Rewrites resumes without inventing achievements or turning team outcomes into individual ownership.
- Guides DOCX/PDF generation and visual QA when the host agent supports document tools.
- Researches employer viability and the value of the specific role using dated public evidence.
- Compares career paths and offers by role value, company quality, growth, return, risk, and exit options.
- Download or clone this repository.
- Copy
skill/career-application-advisorinto your agent's skills directory. - Restart or refresh the agent.
- Invoke it explicitly:
Use $career-application-advisor to review my resume. Diagnose it first and ask for missing evidence before rewriting.
If your agent does not support Codex-style skills, ask it to read SKILL.md and the relevant files under references/ before performing the task.
Use $career-application-advisor to review this resume. Do not rewrite yet; give me the missing-information questions first.
Use $career-application-advisor to compare my resume with this job description and create a truthful targeted version.
Use $career-application-advisor to research this employer and role. Separate confirmed facts, inference, unknowns, and conflicts.
Use $career-application-advisor to compare these offers against my two-to-three-year career goal.
skill/career-application-advisor/
├── SKILL.md
├── agents/openai.yaml
├── assets/job-seeker-intake.md
├── references/
│ ├── resume-review.md
│ ├── company-research.md
│ ├── career-decisions.md
│ └── privacy-and-safety.md
└── scripts/extract_resume_text.py
The core workflow is model-agnostic. Optional local resume extraction requires Python plus:
pdfplumberfor PDFpython-docxfor DOCX
Creating and visually verifying final DOCX/PDF files depends on the host agent's document capabilities.
Provide your resume, target job description, employer information, or offer details directly to your own agent when they are needed for the task. Data handling depends on the agent platform and environment you use. The skill itself does not automatically publish or add these materials to a shared knowledge base. It forbids invented experience and keeps individual contribution distinct from team or platform outcomes.
Version 1.0.0 does not ship fixed visual resume templates. Detailed risk verification, including positive and potential-risk signals, and market-based salary negotiation are planned for version 2.0.0. See ROADMAP.md.
MIT. See LICENSE.