🤖 AI Agents & Tool Use
LLM and agentic AI security, prompt injection, model threats, AI red teaming and AI governance.
Not a-Muse-d! As Meta bets big on Muse AI agent, early safety mishaps surface
The Straits Times
Strike Graph Launches Integration AI to Automate Continuous Compliance Evidence Using Any Existing Tool
New agentic AI solution leverages APIs to collect compliance evidence from systems and applies it to over 30 frameworks with no duplicate work.

Ubuntu AI security patches are here, if you’re willing to test them yourself
Canonical is letting Ubuntu Pro customers install security patches written by AI agents before its engineers finish reviewing and testing them. The…

Microsoft Takes Perception Agentic Security Platform Out of Preview
Four blue team agents and the Recon Agent are now generally available, while integrations and vulnerability shielding are in preview.

OTCC urges US policymakers to bolster OT cybersecurity, fund resilience, govern AI agents as threats accelerate
The Operational Technology Cybersecurity Coalition (OTCC) urges U.S. policymakers to make critical infrastructure resilience a central pillar of...
Eight in Ten LLMs Leak System-Prompt Secrets: Nevan, an Agentic Capture-the-Flag Study
arXiv:2611.21956v1 Announce Type: new Abstract: We audit instruction-hierarchy enforcement across 363 text-in/text-out models from the OpenRouter catalog. The research question we answer with this…
GuardAgent: A Memory-Augmented Agentic Framework for Adaptive Detection of Adversarial Promotions in LLM-Based Recommendation
arXiv:2611.22274v1 Announce Type: cross Abstract: In traditional recommender systems (RSs), textual information typically serves as auxiliary signals for representation learning. In contrast, in…
CyBiasBench: Benchmarking Bias in LLM Agents for Cyber-Attack Scenarios
arXiv:2605.07830v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly deployed as autonomous agents in offensive cybersecurity. In this paper, we reveal an interesting…
IH-Benchmark: A Conflict-Centered Benchmark for Instruction-Hierarchy Robustness in Large Language Models
arXiv:2607.25987v3 Announce Type: replace Abstract: When a language model receives conflicting instructions from different priority levels, which one does it actually follow? This question lies at…
Signed but Not Safe: Requirement Integrity in Multi-Agent LLM Handoffs
arXiv:2611.07483v2 Announce Type: replace Abstract: Enterprise AI agents split requests across subagents, and each handoff can drop, alter or ignore a requirement. A preregistered study on synthetic…
Risky Business #860 -- Google: Agentic hacking has “crossed a threshold”
On this week’s show Patrick Gray and James Wilson are joined by special guest co-host Jason Kikta, CISO at METR.org, to talk through the week’s cybersecurity news, including: Website management…

Linx Security extends identity governance to AI agent actions
Linx Security has announced four new capabilities that bring every agent under identity control, managed or unmanaged, in any gateway. Shadow AI Agent Detection finds the agents no one sanctioned…
Gemini test proves once again that amoral AI agents will do almost anything if told ‘NEVER give up’
If you design a test to find misalignment, you will find misalignment

Detectify launches Agentic Tooling to verify vulnerabilities in live applications
Detectify has launched Agentic Tooling. It lets organizations’ own AI agents, running in their own environments, proactively investigate suspected vulnerabilities and test whether they can be…

Airlock Digital blocks unauthorized AI agent actions at the endpoint
Airlock Digital has announced the availability of Agentic Control with the release of Airlock Digital v7.1. Agentic Control includes dedicated Agentic Steering capabilities for AI agents, allowing…

Chinese state group developing automated attack tool for cyber espionage, Google says
Google released its latest AI Threat Tracker report on Monday, writing that threat actors are rapidly evolving their use of artificial intelligence in attacks — now leaning more heavily on autonomous…