Week 11: Prompt Injection and LLM Security
Dates: Mar 22-26 · Reading: Handout 9: LLM Security and Prompt Injection
Learning Objectives
- Explain how large language models work and their role in business
- Describe prompt injection attacks and real-world examples
- Identify defensive prompting techniques and guardrails
- Understand LLM security best practices
Monday Session
LLMs and their role in business and security. Prompt injection: manipulating LLM behavior through malicious inputs. Real examples: jailbreaks, prompt leakage, and indirect injection via documents.
Wednesday Session
Defensive prompting: system prompts, guardrails, and output validation. LLM security best practices. In-class CTF mini-event using Hacker101 or CTFd (ungraded).
Lab
Lab 9: CTF Challenge. A beginner-friendly Capture the Flag challenge on Hacker101 or the course CTFd instance. Purely educational; no attacks on external systems.
Quiz / This Week
No quiz this week, Lab 9 is due Wednesday.