Engrain by ByteRush

AGT.10 3 concepts

LLM security

Build LLM systems that resist the OWASP LLM risks.

  1. AGT.10.01OWASP Top 10 for LLM applicationsCoreMid-levelBeing written
  2. AGT.10.02Prompt injection in depthCoreMid-levelBeing written
  3. AGT.10.03Data protection, sandboxing and red teamingSeniorBeing written

Interview questions in this module

  1. What is prompt injection? Why can't you fix it with a better system prompt?
  2. How would you secure an agent that reads emails and can send messages?
  3. What is excessive agency?
  4. How do you prevent an LLM app from leaking sensitive data?
  5. Walk through the OWASP Top 10 for LLMs.