Inspiration: In Taiwan, “fake buyer” scams target people selling used items on Facebook. The scammer asks the seller to set up a 賣貨便 listing, claims the order is “frozen,” and sends a fake customer-service link that leads to the seller’s online banking. Police recorded 6,735 online shopping fraud reports in January and February 2026 alone, with 賣貨便 the most impersonated platform. Most victims only realize after the money is gone, so we wanted a check they can do before they click.
What it does: A one-minute questionnaire that estimates the risk of a Facebook Marketplace chat partner. It checks profile signals (friend count, account age, fake-looking name or photo) and behavior signals (refusing to meet, external links, moving to LINE, asking for bank details). It then gives a risk score and clear next steps, including calling 165. It’s bilingual (English / 中文).
How we built it: The scoring is rule-based and transparent. We built the checklist from Taiwan’s 165 Anti-Fraud resources, news reports, and community posts on Dcard and Cofacts.
Challenges we ran into: Scam scripts change constantly. Profile signals alone are unreliable because scammers hijack real accounts. 賣貨便 is legitimate, so the question had to target the real red flag. A low score must never feel like a guarantee of safety. Accomplishments that we’re proud of A tool grounded in real Taiwanese scam patterns, not generic advice Simple enough to finish mid-conversation, under pressure
What we learned: Scams follow a predictable pattern: move you off-platform, create urgency, then make you operate your own bank account. Spotting that pattern is more reliable than judging a profile. One critical red flag can matter more than a high total score.
What’s next for Can I Trust This Person? 這人可信嗎: Hard-stop rules for critical red flags A buyer-side mode for fake sellers
Log in or sign up for Devpost to join the conversation.