TECH NEWS

‘Context Bombs’ Cut AI Attack Success by Misleading Agents With Fake Targets

London- and New York-based Tracebit has developed a military style solution to AI attacks. Just as landmines are used to destroy or slow an advancing army, Tracebit has developed AI ‘context bombs’ that are planted in the battlefield. Each bomb comprises crafted and hidden prompts that are automatically ingested by the attacking AI when it crawls the target. These new prompts interfere with the original, possibly malign, purpose of the attacking agent.

Tracebit’s new research (Making attacking agents pay’) extends earlier research that used different context bombs to trigger guardrails. It now focuses on misdirecting and entrapping the agents with fake targets and deceptively complex problems. To achieve this, Tracebit developed nine new context techniques (ingested fake prompts) that stopped all GPT 5.6 Cyber, Kimi K3, Qwen 3.8 Max, and DeepSeek V4 Pro attacks. The same bombs also worked against Frontier Models, cutting GPT-6 Astra’s successes by 54%...

Copyright of this story solely belongs to www.securityweek.com. To see the full text click HERE

https://assets.newatlas.com/dims4/default/af04926/2147483647/strip/true/crop/6513x3419+0+464/resize/1200x630!/quality/90/?url=https%3A%2F%2Fnewatlas-brightspot.s3.ap-southeast-2.amazonaws....

Harvard smartphone app guides the blind faster, with fewer collisions

If you have standard human vision (even if assisted by glasses), chances are you’re adept at finding things, people, and paths to them. But if you have impaired vision, or none at all, countless tasks that are simple for fully sighted people become difficult, dangerous, or even impossible. Fortunately,

https://www.microsoft.com/en-us/security/blog/wp-content/uploads/2026/10/MSFT-NVIDIA-SCI.jpg

Managing open-weight model risk in enterprise deployments

The growing AI landscape is giving enterprises more choice in how they build, customize, and deploy increasingly capable models. Open-weight models are an important part of that landscape. They give enterprises access to model weights and greater flexibility to inspect, evaluate, modify, and run models within their own environments. While

https://devblogs.microsoft.com/foundry/wp-content/uploads/sites/89/2026/10/Microsoft-Copilot-harness-DevBlog-header-4-1.webp

Introducing the Microsoft Copilot harness in Foundry Agent Service

A model can reason about a task. Completing that task takes more: the right context, tools, execution environment, and controls, and an agent loop that brings them together. Microsoft’s unified harness strategy brings a shared execution foundation across our agent experiences while preserving customer choice. In Microsoft Foundry, that

https://azure.microsoft.com/en-us/blog/wp-content/uploads/2026/10/azure-cloud-hsm-audit-protection-header.png

Azure Cloud HSM Hardware Audit Logging: Verifiable audit trails for security and compliance

Organizations running regulated and security-sensitive workloads need trusted evidence of which authenticated users performed management operations on their hardware security modules (HSMs), what actions they took, and when. Today, we are announcing Hardware Audit Logging for Azure Cloud HSM, which collects audit records generated by the HSM itself for supported

Read more

Harvard smartphone app guides the blind faster, with fewer collisions

Managing open-weight model risk in enterprise deployments

Introducing the Microsoft Copilot harness in Foundry Agent Service

Azure Cloud HSM Hardware Audit Logging: Verifiable audit trails for security and compliance