CareerOps is an open-source job-search system that runs locally inside an AI coding CLI (Claude Code, OpenCode, Codex, Antigravity CLI, Grok Build CLI, Pi, Qwen, Kimi, GitHub Copilot) through one shared skill built on the open agent-skill standard. You paste a job URL or description, and it checks that the posting is still live, scores the role against your CV, and, when the fit is there, drafts a tailored CV, cover letter, and application answers that you review and submit yourself. All data stays on the machine, and the only outbound traffic is to the model provider you chose.
Evaluation runs through blocks A to H: a role summary, a CV match that weights each requirement and labels whether the weight came from the job description's wording, its structure, or an estimate, level strategy, compensation research, personalization, interview prep in STAR+R form, and a posting-legitimacy check. The score lands on a 1.0–5.0 scale, and the tool advises against applying below 4.0, which you can override. A code check blocks a PDF whose numbers or facts appear in neither the CV nor the article digest, as a guard against fabrication.
Around the core loop sit a portal scanner (100+ pre-configured companies and 35+ search queries across Ashby, Greenhouse, Lever, and Wellfound, 55+ provider modules, and an optional Playwright liveness check), batch evaluation with headless CLI workers, a Go terminal dashboard for the pipeline, company research and contact discovery with drafted outreach, interview planning, practice, and debriefs, offer and contract reading, follow-up cadences, reply classification, rejection-pattern analysis, and an opt-in plugin system (Gmail, Notion, Apify). It never submits an application, sends an email, or collects telemetry.
Features
- A-H evaluation: role summary, weighted CV match with labeled weight provenance, level strategy, comp research, personalization, STAR+R interview prep, and a posting-legitimacy check, scored 1.0–5.0
- Human in the loop: evaluates and drafts, never submits a form, sends an email, or clicks, and advises against applying below 4.0/5 with an override
- ATS PDF generation: CVs tailored to each job description from your own experience through an HTML template and Playwright, plus Markdown and LaTeX exports
- Cover letters and emails: research-backed letters with four angle prompts and an in-chat approval gate, and formal application email drafts with a subject line and attachment checklist
- Portal scanner: 100+ pre-configured companies and 35+ queries across Ashby, Greenhouse, Lever, and Wellfound, 55+ provider modules, zero-token discovery, an optional Playwright liveness check, and repost detection
- Application assistant: reads Greenhouse, Ashby, and Lever forms and drafts every open-ended answer from the CV and job description, paste-ready
- Research and outreach: deep company research, hiring-manager and recruiter discovery with a drafted LinkedIn note, and funded-company discovery from public feeds
- Interview suite: a bank of STAR+Reflection stories, time-blocked prep plans, practice sessions with feedback, post-interview debriefs, employer red-flag detection, and negotiation scripts
- Pipeline and tracking: a Markdown-table tracker, a Go terminal dashboard with filters and sorting, dedup and health checks, follow-up cadence, reply classification, outcomes, and a calibration report
- Batch processing: parallel evaluation with headless CLI workers
- Multi-CLI: one shared skill for Claude Code, OpenCode, Codex, Antigravity CLI, Grok Build CLI, Pi, Qwen, Kimi, and GitHub Copilot, plus a standalone Gemini API script
- Plugins: opt-in Gmail, Notion, and Apify integrations and a community registry, disabled by default
