secure-ai-job-search

An AI-assisted job search framework with a strict credential and browser-session privacy boundary.

Secure AI Job Search runs on your machine and helps you search, evaluate, rank, and prepare job applications, including login-gated job portals. Credentials, passwords, cookies, session tokens and password-manager vault contents are deliberately kept out of model context and repository files. Your CV, profile and job-application content may be processed by the AI model runtime/provider you choose: if you use a cloud provider, review that provider's privacy and data-retention policy, and use a compatible local model runtime where supported if you need local-only processing. See the AI provider privacy boundary below.

Based on and substantially extended from Mads Lorentzen's ai-job-search, licensed under the MIT License. See NOTICE.md.

What it does

• Searches job portals through portable CLI skills (LinkedIn, freehire, and

regional boards) and ranks the matches against your profile.

• Evaluates fit with an eligibility gate, a language gate, configurable

location/commute rules, and five scored dimensions.

• Routes application documents by market: Japanese-convention employers get

separate 履歴書 and 職務経歴書; English/international employers get one English resume.

• Reviews drafts independently: a fresh-context, read-only reviewer runs

after drafting and before rendering.