To enhance our project and the next diagram artifact version we added for v1.1, I think we should include information about traditional REST API and websocket architecture into the diagram and how this interoperates with LLM state within Generative AI.
This architecture can be recon'd by an adversary in attempt to perform prompt injection or other techniques to exploit vulnerabilities within the LLM application.
To enhance our project and the next diagram artifact version we added for v1.1, I think we should include information about traditional REST API and websocket architecture into the diagram and how this interoperates with LLM state within Generative AI.
This architecture can be recon'd by an adversary in attempt to perform prompt injection or other techniques to exploit vulnerabilities within the LLM application.
Reference Examples: