Job Search Utility
An automated job application pipeline built on macOS Folder Actions and Claude Code. Drop a job description into a folder — as a PDF, image, DOCX, or URL — and the system extracts the details, researches the company, and generates a tailored cover letter, research notes, and application strategy. Fully autonomous, no manual steps.
Built entirely around Claude Code CLI — macOS Folder Actions trigger Claude to process, research, and generate application materials with zero manual intervention.
Job applications shouldn't be manual labor.
Tailoring a resume, writing a cover letter, and researching a company for every application takes hours. Most of that work follows the same pattern: read the job description, understand what they're looking for, match it to your experience, and write materials that make the connection clear. This system automates the entire pipeline — from extracting text from a PDF to generating a finished cover letter — while maintaining the quality of hand-written applications.
The automation pipeline.
Drop a file, trigger the pipeline
macOS Folder Actions watch the INSERT-JOBS-HERE folder. When a file lands — PDF, image, DOCX, or URL — an AppleScript fires with a 10-second debounce (to wait for multiple files) and a lock file to prevent duplicate runs.
Extract and parse
Claude Code reads the file: OCR for images, text extraction for PDFs/DOCX, web fetch for URLs. It identifies the company name, job title, requirements, and key qualifications from the raw text.
Company research
Claude searches the web for the company: overview, recent news, culture, relevant people to contact on LinkedIn, and anything that would strengthen the application. All research is compiled into structured notes.
Generate tailored materials
Using about-me.md (background, skills, emphasis table), resume-master.pdf, cover letter template, and 2-3 recent applications for tone reference, Claude generates a tailored cover letter, detailed research notes with fit analysis, and application strategy — all saved to a dated output folder.
Learn from past applications
Running 'claude update context' scans completed applications in previous-applications/ and updates about-me.md with claims that have been emphasized across multiple applications, building a stronger profile over time.
Zero manual steps.
From the moment a file hits the folder to the moment finished materials appear in the output directory, no human interaction is required. The AppleScript handles file detection, Claude Code handles processing, and the folder structure handles organization. Even cleanup is automated — processed files move to processed-jobs/, originals archive to the output folder.
Folder-driven automation.
A file-system-based pipeline where folder structure is the interface and Claude Code is the engine.