career-ops

career-ops (also written as careerops) is an open-source, locally run AI job-search system designed to be used from an AI coding CLI such as Claude Code, Codex, OpenCode, Antigravity CLI, Grok Build CLI, or Qwen. It is not a one-click application autofiller; it is positioned as an evaluation and tracking pipeline, with the README explicitly stating that it never submits applications automatically and that the user always has the final call. Purpose and workflow The core flow starts when the user pastes a job URL or job description. The system classifies the role into an archetype (for example LLMOps, Agentic, PM, sales, forward-deployed engineering, or transformation), then runs an A-H evaluation that produces: a role summary, CV match analysis with per-requirement importance and source attribution, level strategy, compensation research, personalization notes, STAR interview stories, a posting-legitimacy check for scams or ghost jobs, a work-authorization blocker signal for no-sponsorship JDs, and (for high-scoring roles only) a draft offer. Each evaluation yields a structured markdown report, a tailored ATS-optimized PDF CV, and a tracker entry. Key features evidenced in the README: - Auto-pipeline: paste a URL or JD and get evaluation, PDF, and tracker entry. - A-H evaluation reports: scored 1-5 by holistic judgment across five dimensions, with separate score-neutral signals for requirement importance and posting legitimacy. Block H (offer draft) is only produced at scores of 4.5 or above. - Tailored CV PDF generation: ATS-oriented, keyword-injected, with a defined visual design using Space Grotesk and DM Sans. - Cover letter generator: research-backed drafts with keyword mirroring, angle prompts, and an approval gate before generation. - Application email drafts: formal recruiter/referral/cold emails with subject line, attachment checklist, and source-backed fit points. Draft-only; the tool sends nothing. - Portal scanner: more than 100 pre-configured companies across the AI ecosystem and 45+ search queries, with provider modules for ATS APIs, board-wide feeds, XML/RSS feeds, and local parsers. - Funded company discovery: a command to surface recently funded companies from structured public feeds. - Batch processing: parallel evaluation using headless CLI workers. - Terminal dashboard: a TUI to browse, filter, and sort the application pipeline. - Interview suite: time-blocked prep plans, practice sessions with feedback, post-interview debriefs, and a company red-flag detector. - Offer stage support: contract reading companion, lawyer question list, and salary-gap analysis between desired, advertised, and actual compensation. - Follow-up and reply tools: cadence calculator, seeded reminders, and employer-reply classification. - Pattern analysis: rejection patterns, per-ATS-channel advance rates, funnel stats, and repost/ghost-job detection. - Plugin system: opt-in integrations such as Gmail, Notion, and Apify, disabled by default. - Company research and contact discovery: deep research prompts on company strategy and culture, plus identification of hiring managers/recruiters/peers and drafting of short LinkedIn messages. Usage model The tool is designed to be adapted by the AI coding CLI itself. Modes, archetypes, scoring weights, negotiation scripts, and profiles are plain files that the assistant can read and edit, and setup is conversational: on first launch the system asks about the user's CV, profile, and target roles. Supported invocation includes slash commands like /career-ops (with subcommands for pipeline, scan, pdf, cover, email, batch, tracker, apply, outcome, contacto, deep, training, and project) and plain-language prompts for CLIs that do not register slash commands, such as Codex. A standalone Gemini API script is also included for evaluation without installing a full CLI. Setup options Quick start is one command: npx @santifer/career-ops init, which clones the latest release and installs dependencies. A global npm install is also available for direct use of the career-ops binary. Manual git clone setup is documented, including dependency installation, Playwright Chromium for PDF generation, a doctor command to validate prerequisites, profile/portal YAML configuration, and a markdown CV file. The README notes that first evaluations will be weaker until the user feeds the system context such as CV, career story, proof points, and preferences. Licensing and provenance The project is MIT licensed and includes a trademark policy. It was created and maintained by Santiago Fernández de Valderrama Aparicio, who reports using it to evaluate 740+ job offers, generate 100+ tailored CVs, and land a Head of Applied AI role. The README also links to a HIRED.md wall of public verification stories and a case study. The project refers to itself as the first reference implementation of the "CareerOps Manifesto," a document users can read and sign. Context and caveats This summary reflects only what is stated in the README. The claims about verified hires, press coverage, star growth, and personal outcomes are the author's assertions and were not independently verified here. The important design constraint repeated throughout the documentation is human-in-the-loop control: AI evaluates and recommends, the user decides and acts, and the system never automatically applies or sends messages.