Sign in Play free
Free course · no login

Prompt injection in 6 short lessons

A free, no-login course that explains prompt injection in plain language, lets you practice the core moves, and ends with practical controls for safer AI applications.

Start the course
  1. 1What prompt injection is (and why AI apps fall for it)
  2. 2Pretending to be someone
  3. 3Telling a story
  4. 4One small step at a time
  5. 5When the attack hides in the data
  6. 6When AI can act: tools, agents, and how to defend
  1. 01 What prompt injection is (and why AI apps fall for it)What prompt injection is (and why AI apps fall for it), in short: an AI app reads its own instructions and your message as one stream of text, so a well-written message can change what the app does. No hacking tools and no code, just words the app was never meant to obey. 5 minLive challenge
  2. 02 Pretending to be someonePretending to be someone works when an AI application mistakes a role claim for permission. Saying “I am the auditor” can sound relevant and urgent, but it does not prove identity, scope, or purpose. The model may still disclose a protected field unless another system enforces access. 5 minLive challenge
  3. 03 Telling a storyTelling a story can disguise an instruction as fiction, roleplay, or a hypothetical scene. Narrative framing changes how a request feels, not what the application is allowed to reveal or do. If a model treats the story as a new permission, an ordinary creative feature becomes an injection surface. 5 minLive challenge
  4. 04 One small step at a timeOne small step at a time describes a crescendo attack: harmless turns establish a cooperative frame before the user asks for a sensitive step. The final request can look like a natural continuation, so every consequential turn needs a fresh permission check instead of relying on conversational momentum. 5 minLive challenge
  5. 05 When the attack hides in the dataWhen the attack hides in the data, the user may ask a perfectly reasonable question while a document, page, email, or search result supplies the hostile instruction. Indirect prompt injection is a provenance problem: content the model reads is data, even when it contains imperative language. 5 min
  6. 06 When AI can act: tools, agents, and how to defendWhen AI can act: tools, agents, and how to defend, in short: once an AI can read files, send email or change records, a prompt injection stops being a strange reply and becomes a real action. Harmless tools can chain into a harmful path, so give agents the least access they need and a human checkpoint before anything risky. 5 min

Who this is for