About AIAgentThreats

AIAgentThreats is an independent publication focused on the security of autonomous and agentic AI systems.

We cover news, technical analysis, regulation and policy developments, practical guides, and a living glossary of terms used in AI agent security.

What we cover

• Runtime risks of AI agents (goal hijack, tool misuse, memory poisoning, excessive agency)
• Identity, permissions and non-human identity management for agents
• MCP, tool and supply-chain risks
• Guardrails, runtime protection and monitoring approaches
• Standards, frameworks and regulatory developments
• Vendor and solution landscape (editorial directory)

Our bylines are pen names

All content on AIAgentThreats is written and edited by our editorial team and published under disclosed pen names, one per beat. These names are not real individuals, and we do not attach credentials, photos or biographies to them.

• Alex Rivera: pen name for news and the newsletter
• Jordan Hale: pen name for regulation and policy
• Sam Ortega: pen name for analysis and guides
• Riley Quinn: pen name for the glossary and Q&A

No real person is presented as editor of the site. Accountability for every article rests with the publication as a whole.

How we work

Claims are linked to primary sources such as standards documents, CVE records, vendor advisories and official documentation. Editorial analysis is clearly labeled as analysis. Our full policies are set out in our Editorial Standards.

Contact

Corrections, tips and business inquiries: info@aiagentthreats.com