Resposta completa pendente de tradução. Ler a resposta completa em inglês.

How do I defend a large language model application from attack?

Treat the LLM as untrusted code in a sandbox. Enforce strict input validation, scope every tool call to least privilege, separate the user-supplied context from the system prompt, instrument prompt-injection detection, rate-limit aggressively, and run an AI red team against the system before launch. The OWASP LLM Top 10 is the reference checklist.

Ver resposta completa em inglês →