← Back to library

Let a terminal agent hunt roles and fill applications

Artur Erofeev drove his own logged-in Chrome from Claude Code over remote debugging. A skill file covers two modes: search LinkedIn from the résumé and write matches to a CSV, or read each posting, write a tailored cover letter, fill the form and submit. It handles LinkedIn quick-apply and common applicant-tracking forms, and logs each submission.

  • OutcomeTurn your résumé into a job shortlist and submitted applications with matching cover letters, while unsuitable roles are skipped.
  • SetupAn afternoon+
  • LevelIntermediate
  • CreditArtur Erofeev
View original
Works with
Claude Code

Published Oct 23, 2026Updated Oct 23, 2026AI dry run on sample input: passed Oct 23, 2026Source verified Oct 23, 2026

When should you use it?

  • You are applying to many similar roles each week
  • Forms keep asking the same questions
  • You want a log of everything sent

Not for: Senior searches built on referrals; Sites whose terms forbid automated applications

What do you need, and what do you get?

You need

  • A résumé in Markdown and PDF
  • A cover-letter style template
  • Chrome started with a debugging port and a dedicated profile

You get

  • A deduplicated CSV of matching postings
  • Role-specific letters written in your voice
  • A submitted-jobs log

How does it work, step by step?

  1. Start a dedicated browser profileOpen Chrome with its debugging port on and a separate profile where you are already signed in to job sites.
  2. Lay out the filesKeep a Markdown résumé, a cover-letter style guide, a to-apply list and a submitted log in one folder.
  3. Write a two-mode skillOne mode searches and collects matches; the other reads a posting, writes the letter and fills the form.
  4. Let it skip poor fitsTell the agent to pass on roles that do not match your background instead of applying everywhere.
  5. Save per-site scriptsStore the form-filling scripts it writes for each applicant system so later runs are faster.

The prompt

Our own prompt, inspired by Artur Erofeev and written to work in Claude Code. Values in lime are placeholders.

Prompt0 variables
Mode: [find | apply]. Find: read my résumé in CV_related, search LinkedIn for [roles] in [locations], and record only strong matches in my shortlist file with company, title, link and a one-line reason. Skip duplicates. Apply: for each approved row, read the posting, write a cover letter using my template's tone, fill the form using only facts from my résumé, and pause before submitting if a question is not covered. Log every submission.

Inspired by Artur Erofeev: How to Automate Job Applications with AI (Full Setup Guide). The prompt above is written by playbooks.bot in our own words; nothing from the original is quoted.

What can go wrong?

  • Automated submissions can breach a site's terms; check before using it at volume.
  • Review the first batch of letters closely, since a confident wrong claim reaches a hiring manager.

Our take

Artur Erofeev's setup is useful for a high-volume search, with one firm caveat: quality beats count, so keep the skip rule strict. Run the search mode alone for a few days, prune the list by hand, and only then let apply mode loose on roles you approved. Keep the submitted log, because recruiters will call about openings you no longer remember. Lock answers about work authorization, salary expectations and demographics as fixed text you wrote yourself rather than letting the agent improvise them. Use a dedicated browser profile so the agent does not wander into banking or personal tabs. Check each site's rules on automated applications before running at scale; account bans are common. Review a sample of cover letters for accuracy and tone before trusting the pipeline. Cap daily submissions so your inbox stays manageable when replies arrive. Track reply rates after a week and tighten the skip rule if most applications go nowhere.

Where the time goes back: Repetitive form fields and letter tweaks are what make volume applying slow; the agent does those while you choose the roles.

FAQ

Does it need paid tools?
The author used only Claude Code usage costs, Chrome and Selenium, with no extensions or subscriptions.
Why not run headless?
Using your real signed-in browser avoids log-in walls, cookies and most captchas.

Source and credit

Inspired by Artur Erofeev. We describe the idea in our own words, link to the original, and wrote the prompt ourselves; nothing is quoted.

Inspired by
How to Automate Job Applications with AI (Full Setup Guide)
Creator
Artur Erofeev
Published on
gosu.team
Original date
Mar 17, 2026
License
Not openly licensed (the poster keeps all rights), so we quote nothing and link to the original
Originally for
Claude Code
AI dry run
Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. PASS: Gates held on desk check; no live sends or purchases.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails.
Verified
Oct 23, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.

Are you the creator? Claim or correct this page.

More in Career & Learning

All 37 Career & Learning playbooks →

Related playbooks

Career & LearningWorkflow✓ AI dry run

Have a coding agent review your Logseq journal

Point a coding agent at your note app's command line tool so it can summarize months of journal pages and report on tagged items such as job applications.

Credit: wosk on Hacker News

Under an hour
Career & LearningSkill✓ AI dry run

Learn any topic with a strict Socratic tutor skill

Load Jane Street's open-source tutoring skill into Claude, ChatGPT or a coding agent so that when you ask to be taught something, it first tests what you already know, keeps a 0 to 5 level for each part of the topic, quizzes you one question at a time without handing over answers, re-tests older material by surprise, and only ends the course after a final mixed exam.

Credit: janestreet

Minutes
Career & LearningWorkflow✓ AI dry run

Let the agent see its own 3D output

Give a coding agent a way to position a camera and take screenshots of the 3D scene it builds, so it can check its own work in a loop.

Credit: neomantra on Hacker News

An afternoon+
Career & LearningWorkflow✓ AI dry run

Research the right people and draft personal cold emails

Get a sourced list of 25 to 30 relevant people with reasons and contact routes, plus drafts you send yourself, in about an hour instead of three.

Credit: 4 the curious

Under an hour