DeepFrontend
Learning Paths
Practice
System Design & DS
Design Studio
Data Structures Curriculum
Real-World Case Studies
Careers
Job Board
Resume Builder
Blog
Sign in
Search
⌘K
Go Pro
Loading compiler resources...
AI for Frontend Developers Quiz Practice
Active Recall Drill Session
← Exit Session
Question 1 of 1
hard
AI for Frontend Developers
What is the primary architectural mitigation for prompt injection via untrusted user content flowing into a prompt?
A
Simply asking the model nicely in the system prompt not to follow injected instructions is sufficient on its own
B
Keep untrusted content clearly demarcated as data (not instructions), and scope tool capabilities to the minimum a feature genuinely needs
C
Prompt injection cannot be mitigated at all
D
Always disable all tools to eliminate the risk
Check Answer