SketchySniffer

Source Code

Front-end: https://github.com/katies-toast/sketchy-sniffer Back-end: https://github.com/max3p/SketchySnifferAPI

What Inspired Us

We built SketchySniffer after noticing a common pattern. When people browse online marketplaces like Kijiji or Facebook Marketplace, they often act on impulse. The excitement of a good deal can override caution.

These platforms include many legitimate sellers, but they also attract scammers and bad actors. In the rush to secure a bargain, users can overlook red flags or ignore the feeling that something seems off.

We are addressing Prompt #1: enhancing critical thinking through problem solving and reflection. Instead of focusing only on detecting scams, we asked a better question. How can we help people pause and think before they act? SketchySniffer is our answer.

What We Learned

Scams are not purely technical problems. They are behavioural and psychological.

Scammers exploit predictable cognitive biases such as urgency, scarcity, anchoring, and authority. A listing that says “must sell today” or shows a dramatic discount is designed to trigger emotional decision making.

We learned that critical thinking is less about intelligence and more about interruption. It is about inserting a small moment of reflection between stimulus and action.

We calculate risk by combining different warning signs. Each warning sign has a score and an importance level. We multiply each score by how important it is, then add everything together to get one final risk score.

In simple terms: Risk score = (each signal × its importance) added together.

But the formula alone is not the solution. The real value comes from combining pattern detection with structured reflection.

How We Built It

SketchySniffer is an AI powered decision reflection tool for high risk online interactions.

Users paste a marketplace listing into our platform. We extract key data such as title, description, and price, then analyze it for known scam patterns, linguistic red flags, and pricing anomalies.

Instead of returning a simple yes or no, we provide:

A scam risk score from 0 to 100
A risk level of low, medium, or high
A clear explanation of the signals detected
A short reflection component to re engage critical thinking

The system is designed to augment judgment, not replace it. The goal is to support better thinking.

Challenges We Faced

Balancing sensitivity was a major challenge. If we flagged too many listings, users would lose trust. If we flagged too few, we would miss real risks.

We also ran into technical hurdles accessing Facebook Marketplace data, since login requirements create both technical and compliance barriers. Because of this, we pivoted to Kijiji.

Finally, we had to control scope. We had many feature ideas, but we focused on delivering a clear, functional MVP aligned with the prompt.

The Outcome

SketchySniffer is a fully built, live tool.

It gives users a practical way to insert a pause between impulse and action. Instead of reacting immediately to a compelling deal, they are encouraged to slow down, review the signals, and think critically.

If it helps even one person avoid a risky transaction by reflecting before acting, then we have succeeded.

This project reinforced a principle we believe strongly in. Technology should not only make decisions faster. It should help people think better.