"Seek, and ye shall find" — Matthew 7:7Building systems that measure and contain the cyber capabilities of frontier AI. Creator of Project Simurgh — a containment-attestation framework that red-teams LLM agents and produces signed, offline-verifiable evidence of what an agent was allowed to do after a guardrail miss: 138/138 classifier-missed cases contained, live-agent attack success cut 9/140 to 0/140 on AgentDojo. Evaluated Claude in Anthropic's safety-evaluation program. Working to give defenders the advantage against advanced-AI cyber risk.
IEEE-format research paper exposing a structural vulnerability in WebRTC-based exam proctoring. 100% evasion on Windows 10/11 and macOS 14–26 using documented OS display APIs. Responsibly disclosed to vendors.
Provider-agnostic verifiable containment-attestation framework for agentic AI, evolved from the defensive counterpart to The Invisible Window research. Produces Ed25519-signed, offline-reproducible evidence of what an AI agent was allowed to do after a guardrail miss — contained 138/138 classifier-missed cases and cut a live agent's attack success from 9/140 to 0/140 on AgentDojo.
Local-first LLM knowledge engine. Ingests any document, extracts structured knowledge (claims, concepts, entities, decisions), and exposes it to AI agents via an MCP stdio server. 218 tests passing.
Privacy-first offline QR & URL security scanner built with Kotlin Multiplatform. 100% offline analysis with 5 platform targets.
AI-native Campus OS transforming university PDF syllabi into structured, agent-readable data. Accepted into the Macquarie University Incubator as a formal startup. Full student operations suite with 503 tests across 92 files, live at syllabus-sync.app.
AI-powered Git client for managing multiple identities and generating semantic commits. Built with Electron and React.
Cyberpunk-styled personal media vault with React frontend, Litestar API, and Supabase auth. AI-assisted recommendations, encrypted takeaways, and hardened cookie-based auth.
Red-team LLM agents and measure the cyber-capability uplift of frontier AI. Characterise safety boundaries, contain guardrail misses with signed, offline-verifiable evidence, and publish reproducible findings.
Defense-in-depth that gives defenders the advantage. Offensive research — exploit classes and agent attack campaigns — becomes verifiable containment and real-world defensive tooling.
Hands-on vulnerability research and AI safety experimentation. Current work: cross-platform exploit development, AI capability uplift measurement, and safety boundary characterisation.