All news with #agent security tag

🛡️ Microsoft Security explains how rising agentic autonomy reorients security from models to how agents are assembled, constrained, and governed inside applications. The post identifies amplified risks—agent hijacking, intent breaking, data leakage, supply chain compromise—and shows why the application layer is decisive because builders fully control permissions, tool access, and failure handling. It recommends concrete design patterns: agents as microservices, least permissions, deterministic human-in-the-loop, and distinct agent identity to limit blast radius and preserve auditability.

🔍 AI coding agents now operate across IDEs, terminals, and extension runtimes, so defenders must expand focus beyond source code to repository files, instruction and runtime settings, and third‑party extensions that shape agent behavior. VirusTotal Code Insight and agentic threat intelligence apply semantic analysis to detect malicious intent in syntactically valid artifacts and link findings to broader campaigns and supply‑chain risks. Examples—weaponized tasks.json, malicious Skill.md, redirected settings.json endpoints, and sabotaged extensions—illustrate how semantics can enable exfiltration, privilege escalation, and stealthy attacker control.

🔐 Agentic AI is already operating across enterprises, executing tasks and taking actions often without meaningful security involvement. Security teams must develop hands‑on fluency — build and test agents, understand integrations like the Model Context Protocol, and enforce scoped configurations — because policy alone won't close the gap. The piece distinguishes three agent classes (productivity, MCP‑connected vendor agents, and custom user agents) and emphasizes configuration, access scoping, and training such as SANS SEC545 to reduce exposure.

🔒 OTT Cybersecurity LLC — the team behind Lyrie.ai — announced acceptance into Anthropic’s Cyber Verification Program and the public release of the Agent Trust Protocol (ATP). ATP is an open cryptographic standard that enables real-time verification of an AI agent’s identity, authorized scope, attestation status, delegation, and revocation. The protocol is royalty-free, slated for IETF submission, and a reference implementation is published under an MIT license. Lyrie positions itself as the security layer for autonomous AI agents operating on the internet.

🔐 Google Cloud announced IAM enhancements at Next '26 to secure agentic AI with Agent Identity and an Agent Gateway. Agent Identity, built on the open SPIFFE standard, provides cryptographically verifiable, automatically provisioned identities and new OAuth/certificate tooling. The Agent Gateway, Identity‑Aware Proxy, and Context‑Aware Access extend Zero Trust to agent traffic, while policy updates, guardrails, and Model Armor add runtime defenses.

🔒 AWS has announced general availability of the AWS MCP Server, a managed endpoint that gives AI coding agents secure, auditable access to AWS services using the Model Context Protocol (MCP). The server is part of the Agent Toolkit for AWS and enforces IAM-based guardrails while emitting CloudWatch metrics and CloudTrail logs so teams keep visibility and control. It supports calling any AWS API through a single tool, sandboxed Python execution for multi-step tasks without filesystem or shell access, and a new agent skills format for on-demand, curated procedures. The service is available at no additional charge; customers pay only for the AWS resources agents consume.

🚀 AWS announced the Agent Toolkit for AWS, a production-ready suite of tools and guidance to help AI coding agents build on AWS with fewer errors, lower token costs, and enterprise-grade security controls. The toolkit includes validated agent skills, a fully managed MCP Server, and easy-to-install agent plugins. At launch AWS is shipping more than 40 skills across infrastructure-as-code, storage, analytics, serverless, containers, and AI, plus three plugin bundles for Core applications, Data Analytics, and Agent development. The Agent Toolkit is available at no additional charge; customers pay only for AWS resources consumed.

🛡️ Analysts and Orchid Security warn that enterprises are deploying AI agents faster than governance can keep up, creating an invisible layer of "identity dark matter" that conventional IAM misses. Orchid Security inspects applications at the binary and configuration layer to discover agents, audit compliance, and locate static credentials. Its Ask Orchid assistant answers natural-language questions about active agents, NIST compliance, and credential risks, then recommends prioritized remediation. This in-application observability aims to close the structural gap in identity visibility and enforce purpose-bound, least-privilege controls.

🤖 Amazon Quick now integrates with New Relic's AI agents, enabling on-call engineers, SREs, and engineering leaders to investigate incidents, run NRQL queries, and generate evidence-backed RCAs directly within the Quick workspace. After connecting to New Relic’s remote model context protocol (MCP) server, users can invoke alert insights, log analysis, transaction diagnostics, and user-impact assessments from a conversational prompt. Quick Flows can automate recurring triage runbooks or escalation steps, and responses are surfaced alongside enterprise knowledge in Spaces for context-aware outcomes. The integration is available in all AWS Regions where Amazon Quick operates.

🤖 Join hands-on developer workshops across North America that teach secure, scalable deployment of agentic AI for enterprises. These sessions are practical, bring-your-laptop labs where Platform, Security, and Data practitioners build end-to-end solutions, including GKE cluster hardening, secure sandboxing, and governed data pipelines. Tracks cover GKE + Data and Data Engineering & Analytics, with guidance from Google experts. Attendees leave with runnable labs and operational best practices to accelerate production adoption.

🤖 Gemma 4, released by Google DeepMind, is a new family of open models optimized for local and mobile deployment. The family emphasizes intelligence per parameter, offering ultra-mobile E2B/E4B sizes, a 31B dense model for local GPUs, and a 26B Mixture-of-Experts variant. The shift to an Apache 2 license plus tools like the Agent Development Kit enables offline agentic workflows and commercial use by developers and startups.

🚀 At Google Cloud Next '26, Firestore introduced expanded agentic AI integrations, built-in full-text search, and deeper MongoDB compatibility aimed at accelerating agent-driven applications. The Enterprise edition’s reimagined query engine adds hundreds of expressive query features, JOINs via subqueries, and pipeline operations. Native connections to AI Studio and third-party coding agents plus preview tools like natural language console querying and Usage Insights simplify building agentic workflows. These capabilities are available now in Firestore Enterprise in both Native and MongoDB compatibility modes.

🔒 Microsoft announces Agent 365 is generally available, offering a unified control plane to observe, govern, and secure AI agents across endpoints, cloud, and SaaS. The release adds discovery of local and cloud agents (including OpenClaw, GitHub Copilot CLI, and Claude Code) and integrates with Intune and Defender for inventory, policy controls, runtime blocking, and alerting. Agent 365 also introduces Windows 365 for Agents, partner integrations, and licensing via Microsoft 365 E7 or standalone at USD 15 per user per month.

🛡️ CISA, in collaboration with the Australian Signals Directorate’s Australian Cyber Security Centre (ASD’s ACSC) and other partners, released guidance to help organizations adopt agentic AI systems safely. The guide identifies key security challenges and risks and offers actionable steps for designing, deploying, and operating these systems. It emphasizes risk management, alignment with existing cybersecurity frameworks, and strengthened oversight to help security teams, developers, and decision-makers implement practical governance and controls.

🔒 CISA, the Australian Signals Directorate’s Australian Cyber Security Centre (ASD ACSC), and U.S. and international partners published Careful Adoption of Agentic Artificial Intelligence Services, a joint guide describing cybersecurity challenges and mitigations for agentic AI. The document warns that agentic AI can expand attack surface, cause privilege creep, produce behavioral misalignment, and obscure event records while offering automation benefits to critical infrastructure and defense sectors. It targets developers, vendors, and operators with actionable recommendations — including avoiding broad or unrestricted access to sensitive data and systems, beginning with low‑risk, non‑sensitive use cases, and explicitly accounting for agentic AI in organizational security models and risk posture.

🔒 Microsoft previewed new Microsoft Defender capabilities within the Agent 365 tooling gateway to give security teams near real-time visibility and control over agentic workflows, using webhook-based evaluation to detect, block, and investigate anomalous agent actions before execution. Separately, Microsoft Defender for Cloud now integrates with GitHub Advanced Security generally available to map code changes to production, prioritize alerts using runtime context, and enable coordinated remediation. A hands-on Microsoft Purview demo demonstrates AI-powered data security investigations across the data estate.

🤖 Agents can now provision Cloudflare resources and complete billing through Stripe Projects, enabling end-to-end deployment without manual dashboard steps. Using a co-designed protocol, an agent can discover available services, create or link a Cloudflare account, and receive API credentials to deploy code and register domains. Stripe supplies a payment token (not raw card data) with a default $100/month cap, and human approval can be requested when needed. Any platform with signed-in users can adopt the same orchestration flow.

🔒 Palo Alto Networks announced its intent to acquire Portkey and integrate Portkey’s AI Gateway into Prisma AIRS to provide a centralized control plane for agentic AI. The combined platform will offer a unified API to thousands of LLMs, an agent registry, semantic routing, caching and runtime protections such as Agent Artifact scanning and automated red teaming. Integration with CyberArk is intended to enforce agent identity and least‑privilege controls. The goal is to enable enterprises to move autonomous workloads from development to production with consistent governance and minimal performance tradeoffs.

🔌 At Google Cloud Next ’26, Google announced that more than 50 Google-managed MCP servers are generally available or in preview, enabling AI agents to connect securely to Google and Google Cloud services without local MCP deployments. The managed endpoints integrate with major agent runtimes and frameworks including Gemini CLI, LangChain, ADK, and others, supporting Resources and Prompts as protocol primitives in addition to Tools. The offering emphasizes enterprise-grade security, governance, and observability through native IAM controls, Model Armor content safety, OpenTelemetry tracing, and Cloud Audit Logs.

🖥️ Amazon Quick is now available as a native desktop preview for macOS and Windows, extending the assistant beyond the browser to leverage local files, OS-level notifications, and native desktop controls. The desktop app can read and work with files on the machine without uploading them, surface action-item, calendar, and message alerts, and automate both browser-based and desktop workflows. Memory, knowledge graph, and agents are shared with the web experience, and the preview supports local Model Context Protocol (MCP) connections for coding agents.