Our Research Projects
All research at Quantum 2x is currently closed. Projects may be made available for licence or released as open source in the future. Enquiries can be directed to our contact page.
Model Package Protocol (MPP)
A licensed protocol for secure, signed, and sandboxed AI tool artifacts, commercialised by Quantum 2x. MPP brings container-like isolation to AI agent tool execution with WASM sandboxing, Ed25519 signing, and fine-grained permission boundaries.
Behaviour Verification Systems
Investigating formal and empirical methods to verify that AI models behave within defined, auditable bounds. Research encompasses constraint specification, runtime monitoring hooks, and post-hoc analysis of model decision traces to detect deviations from intended behaviour.
AI Agent Review & Monitoring
Design and prototyping of systems for continuous review and real-time monitoring of AI agents in operational environments. Covers decision tracing, tool-call auditing, anomaly detection, and dashboards for human oversight of autonomous agent pipelines.
Security Protocols for Agents in Production
Research into security architecture for AI agents deployed in live systems. Areas include sandboxed execution, least-privilege tool access, inter-agent trust hierarchies, prompt-injection defences, and supply-chain integrity for agent ecosystems.
AI Agent Audit Tools
Development of tooling and frameworks enabling compliance teams and operators to audit AI agent behaviour. Outputs include append-only cryptographic audit logs, decision replay systems, and structured reporting aligned with emerging AI governance standards.