mlsecdev/spam-detection-system ? reverse-engineered prompt

Build me a simple Python web app that lets me paste in an SMS or short message and instantly tells me if it looks like spam or normal, with an easy to understand confidence score. I want it to feel fast and friendly, with a clean interface where I can type or paste text, hit analyze, and see the result right away.

If possible, also highlight obvious spam signals like suspicious links, common scam wording, or phone numbers in the message, so it feels a little more informative than just a yes or no label. Please use the trained model and vectorizer files that are already in the project if they’re there, and make sure the app actually runs locally without extra setup beyond installing requirements and starting it.

Keep the wording simple for a normal user, maybe with color coded results and a short explanation of what the prediction means. If you need to, look up current docs online, but keep the final app lightweight and easy to run.