Skip to content
Prompt Injection Attacks and Mitigations
josephthacker.com | blog | #ai-security | #prompt-injection | #llm | #plugin-hijacking | #mitigations | #prompt-firewall | #indirect-injection
Summary
Surveys prompt injection risks for LLM-powered services—plugin hijacking, indirect injection via poisoned content, and polyglot attacks—and outlines mitigations like prompt firewalls.
- Published
- Collected