job-hunter

job-hunter

Self-hosted job-search command center distributed as a Claude Code skill, CLI, and local webapp. Scrapes 11 job boards into a local SQLite pipeline, tracks every application kanban-style, fills ATS forms in shadow or auto mode, and tailors Reactive Resume JSON per JD — without sending PII or session cookies to any model.

Architecture

Layer	Stack
Skill	Claude Code plugin — slash commands for discover, apply, dig, tailor-resume
CLI	Python 3.12 + uv — scrapers, form-fill, salary aggregation
Webapp	FastAPI + HTMX on 127.0.0.1:8765 — kanban tracker, metrics, JD viewer
Storage	SQLite jobs.db mirrored to human-readable tracking.md
Forms	Playwright + YAML adapters (Gupy, Greenhouse, Lever, Workday, Ashby)

Features

• 11 sources out of the box — LinkedIn, Indeed, Glassdoor, Gupy, RemoteOK, Remotive, plus a learn.py adapter that drafts YAML from any unknown form
• Pipeline tracking — discovered → queued → applying → applied → screening → technical → behavioral → offer / rejected / withdrawn
• Match score per job from profile + JD heuristics (seniority, country lock, on-site mismatch)
• Salary aggregator — p25/median/p75 from your own scraped data
• Two-mode form-fill — shadow pauses before submit; auto requires 5 reliability gates + explicit consent
• Resume tailoring — emits Reactive Resume JSON tuned to the JD

PII isolation

Personal data (national ID, phone, address, session cookies) lives in ~/.config/job-hunter/secrets/personal.env at chmod 600. The model is explicitly forbidden to read it via Read tool, cat/head/tail/grep, or any pipe whose output enters the conversation. The CLI loads it via python-dotenv in a child process; the model only sees field schemas. CI runs lint_secret_leaks.py to block any PR that would echo PII into logs or screenshots.