Navigating the future of AI agent security with Dan Moore

Summary

In this episode of the Overcommitted Podcast, Erika and Brittany discuss the evolving landscape of AI agents and their implications for security and identity management. Joined by expert Dan Moore, they explore the challenges posed by non-deterministic agents, the importance of granular permissions, and the need for developers to be aware of security practices as AI technology advances. The conversation also touches on industry standards, the role of developers in navigating these changes, and personal reflections on the future of AI.

Takeaways

• AI agents are changing the landscape of software development.
• Non-deterministic agents present new security challenges.
• Granular permissions are essential for securing AI agents.
• Developers must be aware of security practices in AI.
• Industry standards for AI security are still evolving.
• Separation of concerns can enhance security for agents.
• The role of identity and authorization is critical in AI.
• Business implications of AI agents are significant.
• Developers should stay close to business needs and problem-solving.
• The future of AI will require new skills and awareness.

Links

• Dan Moore on LinkedIn: www.linkedin.com/in/mooreds/
• Dan Moore on Bluesky: https://bsky.app/profile/mooreds.com
• Simon Willison - The Lethal Trifecta: https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/
• FusionAuth: https://fusionauth.io/
• AGNTCY: https://agntcy.org/
• Amazon Bedrock AgentCore: https://aws.amazon.com/bedrock/agentcore/
• FusionAuth Guide to OAuth: https://fusionauth.io/articles/oauth/modern-guide-to-oauth
• MCP and OAuth: https://aaronparecki.com/2025/04/03/15/oauth-for-model-context-protocol
• MCP Specification: https://modelcontextprotocol.io/specification/2025-06-18/basic/authorization

Hosts

• Overcommitted: https://overcommitted.dev
• Brittany Ellich: https://brittanyellich.com
• Eggyhead: https://github.com/eggyhead