Explore topic
AI security
Follow trust boundaries through prompt injection and hidden instructions.
3 published guides
AI security
Jailbreak vs Prompt Injection | A Document Case Study
An uploaded document can contain instructions aimed at the assistant. Follow one fictional case to understand the trust boundary.
3 min read
AI security
WormGPT vs FraudGPT | What the Evidence Really Shows
Separate the history of WormGPT and FraudGPT from seller claims. A familiar name does not verify the tool behind a website.
3 min read
AI security
AI System Prompts | What Hidden Instructions Really Do
Hidden instructions guide an assistant’s behavior. Learn why a claimed prompt leak needs evidence and where real security belongs.
3 min read