FAQs
Common questions about AgentLISA conversations, history, and AI-assisted security guidance.
What is LISA?
LISA is an AI-assisted workspace for discussing smart-contract code and security. It is designed to help people reason about code and possible risks in a conversational interface.
Can LISA make mistakes?
Yes. Language models can misunderstand code, miss relevant context, or produce false positives. Important findings should be reproduced and reviewed with appropriate tests, trusted tools, and security professionals.
What makes a useful prompt?
Describe the behavior you are investigating, include the smallest relevant code sample, and state any important assumptions. Focused prompts make it easier to identify uncertainty and ask precise follow-up questions.
Where can I find earlier conversations?
Signed-in users can use conversation history to return to previous discussions. Reopening a conversation restores its messages so you can review or continue the same line of reasoning.
What should I avoid sharing?
Do not submit private keys, credentials, secrets, or confidential material that you are not authorized to share.
How can I contact the team?
Use the contact page for questions or feedback.