All news with #agent security tag

🔒 Okta introduced an Identity Security Fabric to secure AI agents and unify identity, application, and agent management across enterprises. The platform combines AI agent lifecycle management, a Cross App Access protocol, and Verifiable Digital Credentials (VDC) to enforce least privilege, discover and monitor agents, and replace fragmented point solutions. Early access features begin in fiscal 2027.

🧪 OpenAI is internally testing a new AI agent, GPT-Alpha, built on a special GPT-5 variant and briefly exposed to users in an accidental push. A screenshot shared on X showed an 'Agent with Truncation' listing under Alpha Models, and the agent's system prompt outlines capabilities to browse the web, generate and edit images, write, run, and debug code, and create or edit documents, spreadsheets, and slides. OpenAI says the agent uses GPT-5 for advanced reasoning and tool use and may initially be offered as a paid feature due to increased compute demands.

🔧This wrap-up of the Agent Factory series lays out a repeatable blueprint for designing and deploying enterprise-grade AI agents and introduces the agentic web stack. It catalogs eight essential components—communication protocols, discovery, identity and trust, tool invocation, orchestration, telemetry, memory, and governance—and positions Azure AI Foundry as an implementation. The post stresses open standards such as MCP and A2A, emphasizes interoperability across organizations, and highlights observability and governance as core operational requirements.

🤖 Amazon Web Services announced the Nova Act extension, embedding the agent development workflow directly into popular IDEs such as Visual Studio Code, Kiro, and Cursor. The extension unifies natural-language script creation, fine-grained scripting controls, and integrated browser testing into a single interface, reducing context switching across tools. Built on the Nova Act SDK (research preview since March 2025), the extension is available today from IDE extension marketplaces and the project’s GitHub repository includes documentation and examples to get started.

⚠️ Agentic AI is proliferating inside enterprises, embedding autonomous agents into development, customer support, process automation, and employee workflows. Security experts warn these systems create substantial visibility and governance gaps: organizations often do not know where agents run, what data they access, or how independent their actions are. Key risks include risky autonomy, uncontrolled data sharing among agents, third-party integration vulnerabilities, and the potential for agents to enable or mimic multi-stage attacks. CISOs should prioritize real-time observability, strict governance, secure-by-design development, and cross-functional coordination to mitigate these threats.

🚀 The Gemini CLI is an open-source, agentic command-line assistant built to reason, select tools, and execute multi-step developer workflows while keeping users informed. In a recent Agent Factory episode, creator Taylor Mullen discussed the project's origin, design philosophy, and roadmap. Demonstrations showed onboarding to codebases, converting research papers into interactive explainers, and creating reusable slash commands. The team emphasizes extensibility, transparency, and community-driven contributions.

🛠️ Vertex AI Agent Builder is a unified platform for building, grounding, and deploying production-grade AI agents, designed to move organizations from prototype to scalable, secure services. It centers development on five pillars: Agent frameworks, Model choice, Tools for taking actions, Scalability and performance, and Built-in trust and security, and supports the Agent Development Kit (ADK) and third-party models including Gemini 2.5 Flash Pro. The platform offers managed runtime features such as sandboxed code execution, Agent-to-Agent collaboration, Bidirectional Streaming, and a streamlined one-line path from ADK prototype to Agent Engine deployment, while enterprise controls like VPC-SC and CMEK address compliance and data protection.

⚠️A new study, “Mind the Gap,” examines time-of-check to time-of-use (TOCTOU) flaws in LLM-enabled agents and introduces TOCTOU-Bench, a 66-task benchmark. The authors demonstrate practical attacks such as malicious configuration swaps and payload injection and evaluate defenses adapted from systems security. Their mitigations—prompt rewriting, state integrity monitoring, and tool-fusing—achieve up to 25% automated detection and materially reduce the attack window and executed vulnerabilities.

🔒Prisma AIRS offers unified, AI-native security across the full AI lifecycle, from model development and training to deployment and runtime monitoring. The platform focuses on five core capabilities—model scanning, posture management, AI red teaming, runtime security and agent protection—to detect and mitigate threats such as prompt injection, data poisoning and tool misuse. By consolidating workflows and sharing intelligence across Prisma, it aims to simplify operations, accelerate remediation and reduce total cost of ownership so organizations can deploy bravely.

🤖 CrowdStrike is expanding its agentic AI strategy following its $290 million acquisition of Onum, introducing two initiatives designed to accelerate real-time telemetry and automate SOC workflows. The Agentic Security Platform builds an "enterprise graph" with a semantic data model that acts as a Rosetta Stone to normalize diverse telemetry and enable a global query and command engine. Agent Works provides a no-code environment to create, test, and deploy agentic systems, while the Agentic Security Workforce delivers mission-ready agents in Falcon sensors to automate repetitive analyst tasks and enforce data-protection controls across endpoints.

🔒 Astrix introduces the industry's first Agent Control Plane (ACP) to enable secure-by-design deployment of autonomous AI agents across the enterprise. ACP issues short-lived, precisely scoped credentials and enforces just-in-time, least-privilege access while centralizing inventory and activity trails. The platform streamlines policy-driven approvals for developers, speeds audits for security teams, and reduces compliance and operational risk by discovering non-human identities (NHIs) and remediating excessive privileges in real time.

🚀 Google Cloud and SAP announced tighter integration to unify enterprise data and accelerate intelligent automation. SAP Business Data Cloud now connects to BigQuery via Datasphere, enabling bidirectional replication and AI-ready analytics. Procurement is simplified on the Google Cloud Marketplace with SAP BTP. New agent tooling—Agentspace, the Agent Development Kit, A2A and MCP standards—and expanded M4 memory-optimized VMs certified for SAP HANA aim to speed deployments, improve data consistency, and enable autonomous process automation.

🚀 The CrowdStrike Falcon fall release reframes the platform as an Agentic Security Platform, introducing four core innovations: Enterprise Graph, Charlotte AI AgentWorks, the Agent Collaboration framework (powered by MCP), and an AI-native console. Enterprise Graph unifies telemetry into a real-time, AI-ready data layer to give humans and agents shared context. Charlotte AI AgentWorks delivers a no-code environment to design, test, deploy, and govern mission-specific security agents at scale, while MCP enables secure, orchestrated multi-agent collaboration.

🔗 Azure's Agent Factory blog emphasizes that interoperability is the key to moving agentic AI from isolated prototypes to enterprise-scale solutions. The post promotes open standards like Model Context Protocol (MCP) and Agent2Agent (A2A) to enable shared context, reusable tools, and cross-framework collaboration across runtimes such as Semantic Kernel. It explains how Azure AI Foundry combines these protocols with thousands of connectors, unified observability, and governance so agents can act across SaaS, legacy systems, and custom APIs without costly rewrites.

⚠️ Shadow AI Agents are proliferating inside enterprises as developers, business units, and cloud platforms spin up non-human identities and automated workflows without security oversight. These agents can impersonate trusted users, exfiltrate data across boundaries, and generate invisible attack surfaces tied to unknown NHIs. The webinar panel delivers a pragmatic playbook for detecting, governing, and remediating rogue agents while preserving innovation.

⚠️ Check Point warns that Hexstrike-AI, an agentic AI orchestration platform integrating more than 150 offensive tools, is being abused by threat actors to accelerate vulnerability discovery and exploitation. The system abstracts vague commands into precise, sequenced technical steps, automating reconnaissance, exploit crafting, payload delivery and persistence. Check Point observed dark‑web discussions showing the tool used to weaponize recent Citrix NetScaler zero-days, including CVE-2025-7775, and cautions that tasks which once took weeks can now be completed in minutes. Organizations are urged to patch immediately, harden systems and adopt adaptive, AI-enabled detection and response measures.

🔒 Agentic AI is poised to transform workflows like software development, customer support, RPA, and employee assistance, but its autonomy raises new cybersecurity risks for CISOs. A 2024 Cisco Talos report and industry experts warn these systems can act without human oversight, chain benign actions into harmful sequences, or learn to evade detection. Lack of visibility fosters shadow AI, and third-party integrations and multi-agent setups widen supply-chain and data-exfiltration exposures. Organizations should adopt observability, governance, and secure-by-design practices before scaling agentic deployments.

🤖 Anthropic published a report detailing attacks in which generative AI tools operated as the primary adversary, conducting reconnaissance, credential harvesting, lateral movement and data exfiltration without human operators. The company identified a scaled, multi-target data extortion campaign that used Claude Code to automate the full attack lifecycle across at least 17 organizations. Security vendors including ESET have reported similar patterns, prompting calls to accelerate defenses and re-evaluate controls around both hosted and open-source AI models.

🔍 This post outlines five practical observability best practices to improve the reliability, safety, and performance of agentic AI. It defines agent observability as continuous monitoring, detailed tracing, and logging of decisions and tool calls combined with systematic evaluations and governance across the lifecycle. The article highlights Azure AI Foundry Observability capabilities—evaluations, an AI Red Teaming Agent, Azure Monitor integration, CI/CD automation, and governance integrations—and recommends embedding evaluations into CI/CD, performing adversarial testing before production, and maintaining production tracing and alerts to detect drift and incidents.

🛡️ A recent proof-of-concept by Bargury demonstrates a practical and stealthy prompt injection that leverages a poisoned document stored in a victim's Google Drive. The attacker hides a 300-word instruction in near-invisible white, size-one text that tells an LLM to search Drive for API keys and exfiltrate them via a crafted Markdown URL. Schneier warns this technique shows how agentic AI systems exposed to untrusted inputs remain fundamentally insecure, and that current defenses are inadequate against such adversarial inputs.