Inspiration
Job scams target the exact people this prompt is about: students and new grads chasing their first professional opportunity. The FTC recorded over 132,000 job scam reports in 2025, with $636 million lost (Source: FTC data, 2025). Students who have never received a real job offer can't tell a fake one from a real one.
What it does
Paste any recruiter text, email or DM. SafeHire marks up every scam sign in red pen right on the message, checks its links and domains against Google Safe Browsing, VirusTotal and domain registration records, and gives a plain-English verdict with next steps and a warning you can send to friends. - Cybersecurity Track
How we built it
- Python and Flask backend with one endpoint,
/api/analyze - Regex extraction of URLs, emails and domains from the message
- Google Safe Browsing API for known phishing links, VirusTotal API for domain reputation across 70+ vendors, and RDAP for domain age
- 12 job-scam rules we wrote (chat-app interviews, fake checks, equipment purchases, gift cards and crypto, upfront fees, early SSN requests) and a 0 to 100 risk score
- JavaScript frontend with Rough Notation for the hand-drawn markups
- API keys stay server-side, and results are cached to stay within free-tier limits
Challenges we ran into
-Free-tier rate limits (VirusTotal allows 4 requests per minute), solved with caching
- Keeping API keys out of the browser and out of GitHub
- A slow AI response made us design every API as optional, so the app always gives a verdict even if a service is down
What we learned
How to combine real threat-intelligence APIs with our own detection rules, how to handle API keys safely, and how to design a security tool for people with no security background.
What's next
A browser extension that checks LinkedIn and Handshake messages automatically, and a way for students to report new scams so the rules keep up.
AI use
We used Claude to help write and debug code. SafeHire includes an optional Gemini explanation layer. Verdicts and risk scores come from our own rules and security checks, and our team can explain all of the code.
