Every publication carries its domain, status, date, and author. Status reflects real program maturity: concept, active research, experimental, validated, commercial development, or published.
Foundational AI ResearchAgentic IntelligencePredictive IntelligenceMachine LearningMultimodal IntelligenceComputer VisionRobotics and Physical AIDigital TwinsAI Security and Governance
9 publications
Active ResearchAI Security and Governance
Threat Models for Tool-Using Agents
Publication date
Oct 16, 2026
Author
SkyNet AI Research
An agent that can call tools inherits the attack surface of every tool it can reach. This paper catalogues threat classes specific to tool-using systems, including instruction injection through retrieved content, privilege escalation across tool boundaries, and exfiltration through legitimate channels.
Answers are generated from SkyNet AI’s published research and initiative records only. If the catalogue does not cover a topic, the assistant will say so.
Collaboration
Submit a research or collaboration inquiry.
We review proposals from universities, research organizations, enterprises, and independent researchers. Include the research question, current evidence, and the collaboration model you are proposing.