A technical publication series

Papers

  1. Machine-Speed Adversaries: Red-Team and Blue-Team Agent Swarms Autonomous security operations Adaptive attackers put defensive judgment under pressure. Specialized agents may help, but their advantage has to be measured against the cost and risk of coordination. Incident commanderSpecialist agentsHuman authority
  2. V2UI: A Framework for Governed Voice-Directed Interfaces Voice-directed interfaces Voice can safely construct interfaces when navigation and composition remain separate, host-governed control planes. Navigation planeDeclarative compositionHost authority
  3. Voice as the Orchestrator Layer Voice orchestration Voice works best as a delegation control plane while keyboard and screen remain the precision and review surfaces. Intent captureBounded delegationVisual verification
  4. Building the Discovery Layer for the Agentic Web Agentic resource discovery Agentic discovery can broaden the available options, but policy and protocol authority must narrow them before execution. Capability manifestsHybrid retrievalAuthorization boundary
  5. Keeping Knowledge Alive Knowledge operations Enterprise knowledge stays trustworthy when initial ingestion and continuous regeneration operate as one governed lifecycle. Progressive ingestionEvent regenerationHuman review
  6. Temporal Knowledge Graphs for Agentic Retrieval Agentic retrieval Time-aware retrieval should identify the valid knowledge object before an agent reaches for authoritative content. Temporal validityProvenanceSource authority