Post

If you use "AI agents" (LLMs calling tools in a loop) you need to be aware of the Lethal Trifecta Combine access to private data, exposure to untrusted content and the ability to externally communicate and an attacker can trick the system into stealing your data simonwillison.net/2025/Jun/16/...

simonwillison.netThe lethal trifecta for AI agents: private data, untrusted content, and external communicationIf you are a user of LLM systems that use tools (you can call them “AI agents” if you like) it is critically important that you understand the risk of …

Simon Willison

@simonwillison.net

And yes, this is effectively me trying to get the world to care about prompt injection by trying out a new term for a subset of the problem! I hope this captures the risk to end users in a more visceral way - particularly important now that people are mixing and matching MCP servers themselves

16/6/25, 1:28 pm3 reposts68 likes

Vinkius

@vinkius-mcp-ai.bsky.social

If you are worried about prompt leakage in these agent setups, we have a connector that specifically detects verbatim exfiltration of system prompts using LCS analysis: https://vinkius.com/en/ai-agent-connect/system-prompt-leakage-detector

vinkius.comConnect System Prompt Leakage Detector to ChatGPT and ClaudeDetects verbatim leaks of system prompts within agent outputs using LCS algorithms.

Jukka Niiranen

@jukkan.bsky.social

16 June 2025

Replace the faces of Tom & Jack with any CEO and Simon Willison and it will be perfect.👌