Understanding MCP security implications

Bill Doerrfeld | May 21, 2025

My talk at APISEC|CON 2025 covered agentic AI and MCP security risks and mitigations

Today I presented at APIsec University's APISEC|CON event, sharing my (limited) knowledge about MCP security implications. Since some attendees asked for them, here are my slides:

SLIDES: Understanding MCP Security Implications [PDF]

As I covered on The New Stack recently, researchers have discovered that MCP is not secure by default. It's prone to vulnerabilities such as tool poisoning, rug pulls, tool shadowing, and remote control execution (RCE).


My presentation covered the hype around agentic AI and the excitement around MCP. It then looks at these risks and suggests some mitigations.


It was very helpful for me to put this together, and I'll post the recording of the session once it's out.


 I'm looking forward to closely following autonomous AI, MCP, and related standards, and what all this means for protecting access to underlying APIs. 


Watch: Understanding MCP security risks (recording coming soon)
Nordic APIs ranked #1 API blog on the web
By Bill Doerrfeld June 7, 2025
Nordic APIs, the API-specific blog I edit, was recently ranked the top API blog online by FeedSpot. After ten years managing this presence, I reflect a bit on the journey thus far.
Tips to improve your AI vibe coding
By Bill Doerrfeld June 3, 2025
Developers are realizing that being productive with AI coding assistants takes a lot more than just asking nicely. There's real craft to it.
AI coding is the easy part. Now it's time to focus on production.
By Bill Doerrfeld May 30, 2025
AI coding is the easy part. Now it's time to focus on DevOps to get it into production. In a recent interview for LeadDev's DirectorPlus, Honeycomb's CTO, Charity Majors, shares expert tips on how to accomplish this.
New tools help LLM devs improve training data decisions
By Bill Doerrfeld May 29, 2025
My latest for The New Stack explores new research from Ai2, DataDecide, which helps LLM developers make better decisions around training data.
Knowing when to use AI coding assistants Doerrfeld InfoWorld
By Bill Doerrfeld May 6, 2025
AI coding assistants are a productivity dream in some cases — and a debugging nightmare in others. So, where’s the line?
How semantic caching reduces LLM API calls
By Bill Doerrfeld May 5, 2025
Semantic caching is like typical caching, but for AI. It could eliminate a lot of redundant API calls to LLMs, reducing costs and improving performance.
Using agentic AI for business workflows
By Bill Doerrfeld April 30, 2025
For CIO.com, leading executives shared with me how they're actively utilizing agentic AI to enhance core business workflows.
Making developer productivity metrics actionable LeadDev DirectorPlus
By Bill Doerrfeld April 25, 2025
Developer productivity metrics are useless if they're just sitting in dashboards. So, how can we use them to direct positive, real-world action?
New study reveals what really drives revenue per engineer
By Bill Doerrfeld April 10, 2025
What leads to a higher revenue per engineer? New benchmarking from DX reveals how areas like R&D spend, org size, and growth rate move the needle.
LLMs can now cite their sources
By Bill Doerrfeld April 9, 2025
My latest post on The New Stack reveals how researchers pinpoint the exact sources behind chatbot responses.
More Posts