Nuno Santos

Teofil Bodea, Masanori Misono, Julian Pritzi et al.

AI agents powered by large language models are increasingly deployed as cloud services that autonomously access sensitive data, invoke external tools, and interact with other agents. However, these agents run within a complex multi-party ecosystem, where untrusted components can lead to data leakage, tampering, or unintended behavior. Existing Confidential Virtual Machines (CVMs) provide only per binary protection and offer no guarantees for cross-principal trust, accelerator-level isolation, or supervised agent behavior. We present Omega, a system that enables trusted AI agents by enforcing end-to-end isolation, establishing verifiable trust across all contributing principals, and supervising every external interaction with accountable provenance. Omega builds on Confidential VMs and Confidential GPUs to create a Trusted Agent Platform that hosts many agents within a single CVM using nested isolation. It also provides efficient multi-agent orchestration with cross-principal trust establishment via differential attestation, and a policy specification and enforcement framework that governs data access, tool usage, and inter-agent communication for data protection and regulatory compliance. Implemented on AMD SEV-SNP and NVIDIA H100, Omega fully secures agent state across CVM-GPU, and achieves high performance while enabling high-density, policy-compliant multi-agent deployments at cloud scale.

João Santos, Nuno Santos, David Dias

The proliferation of unreliable and biased information is a significant problem on the Internet. To assess the credibility of the information retrieved from news websites and other sources, users often resort to social platforms looking for confirmation with trustworthy parties. However, users may be faced with considerable obstacles posed by the platform provider, who can prevent access to certain content. This paper presents DClaims, a system that provides a censorship-resistant distributed service for the exchange of information over the Internet using web annotations. DClaims' fully decentralized architecture relies on Inter-Planetary File System (IPFS) and Ethereum blockchain, both of which offer desirable censorship resistant properties. DClaims is implemented as a web annotations browser extension which allows for the classification of news articles, on news websites. From our evaluation of the system, we conclude that a large scale implementation of the system is practical and economically viable.