Sec-ConvAgent is a specialized AI agent module focused on secure, privacy-preserving conversational interactions. Built on the Agentic-OS framework, it encrypts messages and context logs using robust cryptographic protocols, ensuring that sensitive user inputs remain protected during transmission and storage. It integrates with popular LLMs, automatically managing encryption and decryption without developer intervention. By leveraging secure key management, role-based access controls, and end-to-end encryption pipelines, Sec-ConvAgent enables organizations to deploy conversational agents for healthcare, finance, legal, and other regulated domains. Developers can configure secure channels, define encryption policies, and seamlessly integrate secure dialogue flows within their existing Agentic-OS agent orchestration. Sec-ConvAgent bridges the gap between powerful AI-driven conversational interfaces and stringent security requirements.

DelphiMistralAI is an open-source Python toolkit that integrates the powerful Mistral-7B LLM with the Delphi moral reasoning model. It offers both a command-line interface and a RESTful API for delivering reasoned ethical judgments on user-supplied scenarios. Users can deploy the agent locally, customize judgment criteria, and inspect generated rationales for each moral decision. This tool aims to accelerate AI ethics research, educational demonstrations, and safe, explainable decision support systems.