career-ops: An AI Job Search Pipeline That Scores Listings A-F Before You Apply
career-ops turns AI coding CLIs such as Claude Code into a job-search command center, scanning job portals, scoring listings on an A-F rubric, and tailoring CVs.
At a glance
- What is it?
- career-ops turns an AI coding CLI into a job search command center: it scans portals, evaluates listings with a structured A-F rubric into a 1.0-5.0 score, generates tailored CVs, and tracks applications. The core judgement: it is a filter, not a spray-and-pray tool, and it demands significant upfront context to work well.
- Who is it for?
- Adopt career-ops if you are a technical job seeker who already uses an AI coding CLI, you are willing to invest time feeding it your CV, career story, and preferences, and you want a systematic filter that scores hundreds of listings before you spend effort applying. Skip it if you expect a zero-configuration tool that works well on the first run, or if you prefer a simpler spreadsheet-based workflow.
- Can I use it commercially?
- Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
- Is it still maintained?
- Yes. The repository last received commits 2 days ago.
- What is it written in?
- Mainly JavaScript, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 14, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
What career-ops actually solves
The problem is not a lack of job listings. It is the opposite: too many listings, most of them irrelevant, some of them scams or ghost jobs. career-ops addresses this by turning an AI coding CLI into a pipeline that scans job portals, evaluates each listing against your CV and preferences, and produces a structured report with a global score from 1.0 to 5.0. The intended user is a technical professional who already lives in a CLI environment like Claude Code, Codex, OpenCode, or Antigravity, and who is tired of manually triaging job boards. The README is explicit that this is not a spray-and-pray tool. It is a filter. The author claims to have used it to evaluate 740+ job offers, generate 100+ tailored CVs, and land a Head of Applied AI role, but you should treat those numbers as anecdotal, not as a benchmark you can reproduce.
How the evaluation rubric works
The evaluation is not a simple keyword match. The README describes a structured report with blocks A through H. Blocks A through F cover role summary, CV match, level strategy, comp research, personalization, and interview prep (STAR+R). Block G is a separate posting-legitimacy assessment that checks for scams and ghost jobs, and it never affects the global score. Block H is drafted only when the score is 4.5 or above. The global score is reached by holistic judgement across five dimensions, not by an arithmetic formula. That is a deliberate design choice: it prevents a listing with a perfect keyword match but a terrible culture fit from scoring high just because it ticks boxes. The Work-Auth signal is also separate: if a job description explicitly says no sponsorship, that is flagged as a hard blocker. This separation is useful because it keeps the score about fit, while the legitimacy and work-auth flags act as gatekeepers.
The pipeline: from URL to PDF to tracker
The core workflow is the auto-pipeline: you paste a URL, and the system produces a full evaluation, a tailored PDF, and a tracker entry. The system uses Playwright to navigate career pages on Greenhouse, Ashby, Lever, and company pages. It processes listings in batch, evaluating 10 or more offers in parallel with sub-agents. For each listing, it generates an ATS-optimized CV with keyword injection, using a design with Space Grotesk and DM Sans fonts. It also generates a research-backed cover letter with keyword mirroring and four interactive angle prompts: why, problems, approach, and tone. The cover letter has a draft-in-chat approval gate, meaning you must approve the draft before it is rendered to PDF. The Interview Story Bank accumulates STAR+R stories across evaluations, so after several evaluations you have a bank of 5 to 10 master stories that can answer any behavioral question. This is a data accumulation feature that gets more useful the more you use it.
Getting it running in your CLI
The README does not include a quickstart command in the truncated portion, but it states that career-ops runs on any agent-skill-standard CLI, with a list in docs/SUPPORTED_CLIS.md. The primary target is Claude Code, but Codex, OpenCode, and Antigravity are also mentioned. The installation likely involves cloning the repository and running a setup command, but since the README is truncated, you should check the repository's docs folder for exact commands. The configuration is context-heavy: the README warns that the first evaluations will not be great because the system does not know you yet. You need to feed it your CV, your career story, your proof points, your preferences, what you are good at, and what you want to avoid. Think of it as onboarding a new recruiter. This is not a plug-and-play tool. The quality of the output depends directly on the quality of the context you provide.
Limitations and failure modes
The most obvious limitation is the cold-start problem. The README is honest: the first evaluations will not be great. If you feed it nothing, you will get generic, low-value reports. Another limitation is that the system only scans specific portals: Greenhouse, Ashby, Lever, and company pages. If your job search relies on LinkedIn, Indeed, or niche boards, career-ops will not scan them. The README also strongly recommends against applying to anything scoring below 4.0/5, which means the tool is designed to reject most listings. That is fine if you have a large volume of listings, but if you are in a niche field with few openings, the filter may leave you with nothing. The legitimacy check (Block G) flags scams and ghost jobs, but it is a separate assessment that never affects the score, so a listing could score high on fit while still being flagged as suspicious. You must review that flag manually.
Alternatives and how they differ
The most common alternative is a manual spreadsheet workflow: you copy job descriptions into a table, track application status, and write tailored CVs by hand. That approach costs time but gives you full control and no dependency on a specific CLI or on AI reasoning quality. Another alternative is a dedicated job search platform like Simplify or Huntr, which offer browser extensions and dashboards for tracking applications. Those tools typically focus on the tracking and application-filling part, not on deep evaluation with a structured rubric. career-ops differs by embedding the evaluation logic into an agentic CLI, which means the AI reads the job description and your CV, reasons about fit, and produces a nuanced report. The trade-off is that you must trust the AI's reasoning and invest in the initial context. A spreadsheet does not need context, but it also does not tell you whether a posting is a ghost job.
Maintenance, license, and upgrade cost
The project is under active development, with recent releases including career-ops v1.30.0 and web-v0.8.1 as of August 2026. The primary language is JavaScript, and the license is MIT, which means you can use, modify, and distribute it freely, including for commercial purposes, as long as you preserve the copyright notice. The repository includes a TRADEMARK.md file, so while the code is MIT, the name and branding may have separate restrictions. The maintenance cost is on you: because it runs locally in your CLI, you are responsible for updating the tool, keeping Playwright browsers current, and ensuring your API access to the AI CLI works. The README mentions a community Discord and a HIRED.md file where users can share success stories, but those are not technical dependencies. The upgrade cadence appears frequent, so you should expect to pull new releases regularly to get bug fixes and new portal support.
Who should adopt it and what to verify first
Adopt career-ops if you are a job seeker who already uses an AI coding CLI, you have a high volume of listings to triage, and you are willing to spend the first week feeding the system context about yourself. Skip it if you prefer a lightweight tracker or if you rely on job boards that are not supported. Before you commit, verify that your target portals are in the supported list (Greenhouse, Ashby, Lever, company pages), confirm your CLI is in docs/SUPPORTED_CLIS.md, and read the manifest at career-ops.org/manifesto to see if the philosophy aligns with your approach. The system is a filter, not a generator of applications. It will tell you which listings are worth your time, but you still have to review every report and make the final call.
Editorial conclusion
Adopt career-ops if you are a technical job seeker who already uses an AI coding CLI, you are willing to invest time feeding it your CV, career story, and preferences, and you want a systematic filter that scores hundreds of listings before you spend effort applying. Skip it if you expect a zero-configuration tool that works well on the first run, or if you prefer a simpler spreadsheet-based workflow. Before adopting, verify that your target job boards are among the supported portals (Greenhouse, Ashby, Lever, company pages), confirm that your CLI is on the supported list in docs/SUPPORTED_CLIS.md, and be prepared to review every evaluation manually, since the system itself recommends against applying to anything scoring below 4.0/5.
Community notes