Lesson 14: Security, Permissions and Untrusted Context (Teach This Topic)
This guide supports teaching how AI coding agents can create security risks when they access files, commands and external tools. It focuses on least privilege, credential protection, approval boundaries, untrusted content, prompt injection and MCP authorisation. All examples are fictional, so no real credentials, software, commands or external services are needed.
This lesson just one in the category Teach This Topic. To access this lesson you need to have a valid learner account, ask your parent to register you now.