A Patched MCP Client Can Still Send Its Secret to the Wrong Server
A fixed MCP Python SDK release is only part of the repair for unattended OAuth clients. Pin the issuer, rebind saved registrations and assess possible credential exposure.
AI
Artificial intelligence, local AI, agents, tooling and AI security.
The AI Agent Copied the Attack Into Its Own Reply
In an OpenAI test, a routine email reply carried a prompt injection into the outgoing message. The result was simulated, but the path between agents deserves a real-world test.
Your Salesforce Agent Could Have Sent the Phish in Slack
A malicious CRM lead could cross into Agentforce and emerge as a trusted Slack reply without showing who triggered it. Salesforce has changed the defaults.
The Hidden ChatGPT Task That Read Gmail While the User Saw a Normal Answer
Check Point used a shared internal package service to pass hidden tasks between separate ChatGPT accounts. In its demonstration, the victim saw a normal answer while Gmail data crossed the boundary.
AI Kept Rebuilding This Russia-Linked Malware Until Security Tools Missed It
Anthropic says a Russia-linked espionage operator used AI to monitor detections and keep rebuilding malware until it was no longer detected. The attack cycle, not a single hash, is now the defensive problem.
China Did Not Need the Model Weights. U.S. Agencies Say It Took the Capabilities Through the API.
NSA, CISA and the FBI say six Chinese AI companies distributed millions of requests across providers, accounts and proxies to extract frontier-model capabilities.
The AI Needed Human Approval. So It Invented One.
CLTR’s AI incident data points to a deeper failure: systems that can fabricate human approval and turn oversight into theatre.
Hackers Used AI to Move Faster. Then They Exposed Their Own Playbook.
Two Latin American intrusion campaigns show AI accelerating attacker troubleshooting, but exposed consoles and open directories revealed how the operations worked.
Aurora Used Cursor to Turn Ransomware Into a Repeatable Playbook
Recovered Cursor logs show an Aurora ransomware affiliate using Claude against live victim networks while a skilled human controlled the intrusion.
One Unpatched Langflow Endpoint Could Hand Attackers Your OpenAI and AWS Keys
Attackers are exploiting a critical Langflow endpoint to steal OpenAI keys, AWS credentials and administrator secrets. Patching is only the first step.