A First Line of Defence Against Prompt Injection, in Python
The moment you put an AI chatbot in front of the public, people start trying to talk it into misbehaving. Here's a small Python guard that screens user input for the common prompt-injection tricks before it ever reaches the model…