← Back to work

Job Hunter

A resume-driven job search skill that deep-matches listings on a weighted rubric and discloses coverage gaps instead of hiding them

AgentsProduct View on GitHub →

Sample output

Preferences used: Remote only · target salary $150K–$250K · Titles: Senior Technical Program Manager, Senior Technical Product Manager, AI Program/Product Manager

RankMatch %Job TitleCompanyRemote?SalaryBoard
190%Principal Product Manager (AI Platform & Integrations)NextivaMostly remote$130K–$200KIndeed
288%Senior Technical Program Manager (TPM)SylloRemote (USA)$160K–$190KBuilt In
382%Technical Program Manager, Knowledge SystemsCoinbaseRemote (USA)$149K–$175KBuilt In
480%Senior TPM (Engineering) – AI Tooling & SystemsDeepgramRemote (USA)$152K–$190KBuilt In
580%Senior TPM, Services Tools & DiagnosticsTwilioRemote$188K–$277KIndeed

Trimmed excerpt — full run scored 17 listings across 5 boards. Salary figures are the real range each employer posted; full example in the repo.

The problem

Job board search is noisy by default: hundreds of listings, most a weak fit, and no consistent way to compare a “senior” title at one company against a “staff” title at another. Filtering by hand doesn’t scale, and most job-alert tools match on keywords, not actual fit.

What I built

A skill that parses a resume, asks for work-arrangement and salary preferences up front, searches multiple job boards, and scores every listing against the resume on a weighted rubric before presenting a ranked table. Originally built for my own search, later adopted as one of the broader set of available Claude skills.

How it works

Stack

Claude Code skill, Claude in Chrome (browser automation), fetch-only fallback across six job boards.