itsblurf/codex-job-search-assistant

This project is a job hunting workspace built to run inside OpenAI's Codex tool. Instead of being a normal app you open, it works through a set of named skills that Codex reads and follows, and it is meant to help someone search for jobs, tailor their materials, and keep track of the whole process from one place. A person starts by feeding in real documents such as an old CV, a LinkedIn export, diplomas, reference letters, and notes from past applications. A setup skill reads those documents and builds a profile of the person's background, being careful to mark anything it is unsure about rather than making things up, and asking questions when it finds conflicting facts like mismatched dates or job titles. A second skill can expand that profile by searching the same documents for skills the person may have listed without realizing it, but only if the evidence supports it. Once a profile exists, other skills search supported job sites and save the results, rank saved jobs against the profile with honest strengths and gaps, and generate a tailored CV and cover letter in LaTeX for a specific job. If LaTeX and a PDF text extraction tool are installed, the project can compile the documents to PDF and check whether the text would be readable by an automated applicant tracking system. Every application outcome, such as an interview or rejection, gets logged to a tracker file so the person can review their history and spot which skills keep getting them turned down for jobs, feeding into a final skill that suggests what to learn next. Private documents like CVs and diplomas are kept out of version control by design. The project is aimed at developers or technical job seekers comfortable using Codex, and it is released under the MIT license.