← The Polylog AI Intelligence Brief
Morning Edition · Wednesday, October 28, 2026Published at 2:21 AM EDT · New York
Log-To-Leak Paper Shows Poisoned MCP Tool Descriptions Can Make Agents Leak Data After Finishing Their Task
Across 18 model-server pairs, the attack succeeded 62.64% to 100% of the time while leaving the legitimate task almost untouched, the same week OpenAI and Check Point expanded agent-driven vulnerability research.
A paper posted to arXiv on October 26 by Yuepeng Hu, Chongyu Fan, Sirat Samyoun and Jian Du describes Log-To-Leak, an attack on agents that call tools through the Model Context Protocol (MCP). The attack exploits the natural-language metada…
Continue the AI Intelligence Brief
Track frontier labs, chips, export controls, model releases, regulation, and AI infrastructure.
- 5 AI intelligence signals a day
- Frontier labs, compute, and chips
- Model releases and AI infrastructure
- Source-grounded analysis with confidence labels
The Global Intelligence Brief stays free.
Part of a tracked trend
The Agent Skill Supply Chain Becomes an Attack Surface
As agents load third-party skills, tools and MCP servers at runtime, the gap between what a skill advertises and what it actually does becomes a recurring security failure mode, driving registries, attestation and zero-trust verification into the agent stack the way package signing came to software repositories.
Townsquare
Start a discussion in Townsquare.
More from this edition
- Anthropic's IPO Filing Shows $16.3 Billion in Half-Year Revenue and Hundreds of Billions in Fixed Compute Contracts
- Moonshot Cuts Cyber Capabilities From Kimi K3.1 Pro Before Releasing Its Weights by November 4
- Axiom Says AI Agents Machine-Checked the Classification of Finite Simple Groups in 24 Days
- Google DeepMind Previews Genie 4, a Real-Time World Model That Generates Sound and Dialogue
- EU Governments Back a GDPR Change That Lets Companies Claim Legitimate Interest to Train AI on Personal Data
- OpenAI Pushes GPT-6 Astra Into Oracle's Enterprise Apps While a New Paper Questions Vendor Time-Savings Claims
- Claude Produced 17 Lab-Confirmed Scorpion-Toxin Binders From 318,000 Designs, as Small Clinical Models Approach GPT-4 Scores
Comments
0No comments yet.