career-ops job grader

A new open-source project called career-ops has become one of GitHub’s fastest-rising job-search tools, promising to score roles, tailor resumes, and filter out weak applications. (github.com) As of April 12, 2026, the main repository showed about 29,600 stars and 5,700 forks on GitHub, with recent commits adding features including Codex support, Canva resume generation, and a rejection-pattern detector. (github.com) The repository says the software turns Anthropic’s Claude Code into a “job search command center” with 14 skill modes, a Go dashboard, PDF generation, and batch processing. Its README says it evaluates offers on 10 weighted dimensions and recommends against applying to jobs scored below 4.0 out of 5. (github.com) The basic idea is simple: employers use applicant-tracking systems to sort resumes, and career-ops tries to help candidates pre-screen jobs before they spend time applying. The project says it can compare a user’s curriculum vitae against a job description, generate an applicant-tracking-system-friendly PDF, and log the result in a tracker. (github.com) The tool also automates job discovery. The README says its portal scanner comes preconfigured for more than 45 companies and searches sites including Greenhouse, Lever, Ashby, Wellfound, and company career pages. (github.com) The creator, writing on santifer.io, said the system grew out of a personal search that ran 631 evaluations across 12 modes. The GitHub README gives a larger running total, saying the tool was used to evaluate more than 740 job offers, generate more than 100 tailored curriculum vitae documents, and land a Head of Applied Artificial Intelligence role. (santifer.io) (github.com) The project frames itself as a filter, not an auto-apply bot. Its README says “always review before submitting,” and its legal disclaimer says default prompts tell the model to stop before the final send or apply action. (github.com 1) (github.com 2) That disclaimer also lays out the project’s limits. The maintainer says the code runs locally, sends data only to the user’s chosen model provider, can hallucinate qualifications or company facts, and should not be used to spam employers or violate job-board terms of service. (github.com) The pitch lands at a moment when job seekers and employers are both leaning harder on automation. Career-ops says it uses reasoning about a candidate’s background instead of simple keyword matching, but the repository also warns that early evaluations may be weak until users feed the system more context about their work history and preferences. (github.com) For now, the project’s rise says as much about the hiring market as the software itself: candidates are adopting the same kind of automation that many believe employers already use. Career-ops’ own documentation still keeps a human at the last click. (github.com 1) (github.com 2)