All news with #ai security tag

🛡️ Researchers disclosed a critical issue in Grafana, dubbed GrafanaGhost, that enables zero‑click exfiltration of sensitive telemetry and business data via AI‑powered dashboards. Noma Security reported the chained exploit, which combines indirect prompt injection and a URL validation bypass; Grafana validated the report and released a patch. The attack abuses protocol‑relative URLs and model keywords to trick AI into sending data to attacker servers. Organizations should patch, restrict img‑src, and enforce egress controls.

🔒 The Hacker News is hosting a webinar that examines why identity programs can advance while enterprise risk rises. New Ponemon Institute research finds hundreds of applications remain disconnected from centralized identity, creating an unmanaged dark matter attack surface that AI agents now exploit. Join experts Mike Fitzpatrick and Matt Chiodi for tactical guidance to measure, prioritize, and close identity gaps.

🛡️ The FBI's 2025 Internet Crime Report shows US victims lost more than $17.7 billion to internet-enabled fraud, with the Internet Crime Complaint Center (IC3) receiving over one million complaints in 2025. Cryptocurrency investment scams were the single largest source of financial loss at $7.2 billion, followed by Business Email Compromise and fake tech support schemes. The report also highlights nearly $893 million lost to AI-enabled fraud and 22,364 AI-related complaints, warning that synthetic content and deepfakes are increasingly abused to perpetrate scams.

🔒 Attackers are compressing the time from initial access to lateral movement by using AI, automation and refined TTPs, forcing defenders to adopt prevention-first strategies. The article highlights that average breakout time is about 30 minutes and that exfiltration can sometimes occur in minutes, with extreme cases measured in under ten minutes. It recommends AI-powered XDR/MDR, unified visibility across endpoint, network and cloud, and stronger identity-centric controls to speed detection and response. Automated containment—session termination, host isolation and password reset—should be orchestrated with SIEM and SOAR to reduce dwell time.

🔒 CrowdStrike announced it is a founding member of Project Glasswing, partnering with Anthropic to secure execution of frontier models like Mythos Preview where they run inside enterprises. CrowdStrike emphasizes its sensor-level visibility across endpoints, real-time AI Detection and Response, and Falcon Data Security to govern data and agent behavior at runtime. The company frames deployment governance as distinct from model safety and highlights regulatory and operational requirements for enterprise adoption.

🔒 Microsoft Defender Security Research describes an AI-enabled campaign that abused the OAuth Device Code flow to compromise organizational accounts at scale. Actors used generative AI to craft hyper-personalized lures and automated backend infrastructure (including Railway.com and other PaaS) to generate dynamic device codes at click time, defeating the standard 15-minute expiry. The activity is linked to the PhaaS toolkit EvilToken and shows a marked escalation in automation and scale versus earlier device code phishing campaigns. Post-compromise actions focused on device registration, Microsoft Graph reconnaissance, malicious inbox rules, and email exfiltration.

⚠️ Attackers are increasingly abusing enterprise AI services—poisoning connectors, impersonating Model Context Protocol (MCP) servers, and using platforms as covert C2 channels—to exfiltrate sensitive data and hide malicious traffic. Notable incidents include a counterfeit MCP package siphoning transactional emails, the SesameOp backdoor tunneling commands through the OpenAI Assistants API, and command-injection flaws in Microsoft Copilot and OpenClaw that enabled agent hijacking. Threat actors also automate espionage with Claude Code and assemble modular black‑hat stacks like Xanthorox and Hexstrike. Security teams should treat AI assistants like privileged users, enforce governance, and harden supply-chain and connector integrity.

🔒 This Unit 42 analysis evaluates Amazon Bedrock Agents' multi-agent collaboration from a red-team perspective. The researchers demonstrate a chain of reconnaissance and exploitation—detecting operating mode, enumerating collaborator agents, delivering attacker-controlled payloads, and triggering tool actions—when Bedrock Guardrails and pre-processing are disabled. The report confirms no vulnerabilities in Bedrock itself and emphasizes mitigations such as Bedrock Guardrails, input validation, scoped agent capabilities, and the principle of least privilege.

🔒 A leaked copy of Claude Code has revealed a documented vulnerability that can be triggered when the tool receives more than 50 subcommands. Researchers at Adversa found that subcommands beyond the 50th bypass compute-intensive security analysis and instead elicit a simple user confirmation, creating a risky blind spot. Anthropic has developed a fix — a tree-sitter parser — but it is present only in internal code and not enabled in public builds that customers use.

🔧Envoy is presented as a production-ready data plane for agentic AI networking, arguing that networks must parse protocol payloads and enforce governance centrally rather than acting as blind transports. The post explains how Envoy deframes MCP, A2A, and OpenAI-style traffic to expose protocol attributes to filters and reuse HTTP extensions such as RBAC, ext_authz, and tracing. It also covers per-request buffer controls, session management for streamable transports, AgentCard-based discovery, and integration with control planes for policy rollout.

🔒 The 2026 AI Cybersecurity Summit addresses the widening gap between rapid AI adoption and lagging security by focusing on practical, deployment-stage risk management. Speakers and sessions will explore visibility, governance, and layered protections across GenAI tools, custom models, APIs, and agentic systems. Attendees will receive operational guidance to secure AI as it moves from experimentation to production. The summit emphasizes integrating security, infrastructure, and operations to reduce accumulating risk.

🤖 RSA 2026 highlighted a rapid, industry-wide shift toward AI-driven security, with CISOs clustering into three archetypes—proactive, curious/confused, and blissfully ignorant. Vendors stressed the need to build AI foundations (data/context engines, control planes, execution layers) and then layer agents atop them. Microsoft, legacy security vendors, and AI-native startups all showcased approaches, while pricing, governance, and evolving threats remain open challenges.

🔒AWS outlines four security principles for agentic AI in its NIST CAISI response, arguing existing security frameworks should be extended rather than replaced. It emphasizes secure development lifecycles for both traditional and AI components, continued use of standard controls, and deterministic, infrastructure-level enforcement outside the agent's reasoning ('security box'). AWS applies these through Amazon Bedrock AgentCore, which provides compute isolation, identity and access controls, centralized tool gateways, observability, and secure model execution.

🛡️Open-source components are now central to modern development, but their vulnerability data, maintenance status, and supply-chain integrity are increasingly unreliable. Public vulnerability databases often lack CVSS scores, contain inconsistent metadata, and lag behind exploit availability, leaving teams to guess prioritization. Unmaintained, EOL packages persist across projects, and registries have seen sharp rises in malicious packages and automated worm-like campaigns. AI-assisted coding accelerates development but can amplify these risks by suggesting outdated or hallucinated dependencies and cannot fully remediate legacy or deep dependency flaws on its own.

🔒 The Talos Threat Source newsletter warns that business email compromise (BEC) attacks have been democratised by AI, enabling attackers to cheaply and rapidly craft convincing payment requests that target small community organisations, charities, and businesses. Attackers can automate reconnaissance and generate tailored messages referencing projects, tone, and terminology. Defenders should verify unexpected payment requests via independent channels, enforce procurement controls, and increase awareness. The briefing also flags an automated credential-harvesting campaign exploiting React2Shell in Next.js applications that risks wide-scale token and key theft.

🤖 Microsoft describes a shift from AI as a tool to AI as an embedded attack surface, accelerating tempo, precision, and scale across reconnaissance, malware development, and post-compromise activity. AI-enhanced phishing campaigns now report click-through rates near 54% versus roughly 12% for traditional campaigns, a 450% increase. The blog highlights Tycoon2FA, tied to Storm-1747, as an industrialized, subscription-based phishing ecosystem that automated MFA bypass at scale. Microsoft’s Digital Crimes Unit disrupted the operation, seizing 330 domains with Europol and partners, and urges organizations to prioritize agent inventory, agentic accountability, and lifecycle-integrated intelligence and defenses.

🔍 This article introduces labs demonstrating how to prepare and use data stored in Google Cloud databases to support production-ready AI. It highlights semantic search using embeddings in AlloyDB and Cloud SQL (PostgreSQL and MySQL), multimodal image–text embeddings, and AlloyDB AI functions like on-the-fly semantic evaluation and reranking. It also covers NL2SQL generation via the alloydb_ai_nl extension and points to hands-on modules for moving from tests to production.

🤖 Cloudflare describes how increasing AI crawler traffic—used by retrieval-augmented generation, real-time summarization, and large-scale dataset collection—fundamentally alters CDN cache dynamics. AI agents request high volumes of unique, long‑tail URLs, often in parallel and without shared sessions, producing low reuse and high cache churn that raises misses and origin load. Cloudflare proposes AI-aware caching, traffic filtering, and a dedicated AI cache tier to preserve low-latency human-facing performance while serving diverse AI workloads.

🔐 AI is rapidly reshaping how software is written, deployed, and consumed, pointing toward a future of on-demand "instant software" that is created and discarded as needed. The essay examines how improved AI tools will change the attacker/defender dynamic by automating both vulnerability discovery and, potentially, patch creation. It highlights particularly exposed areas such as IoT and legacy industrial systems and outlines several key unknowns—AI effectiveness on closed-source code, patch reliability, update lag, coordination of defenses, and risks of poisoning or social-engineering attacks. The author sketches optimistic scenarios (self-healing networks, rapid coordinated patching) while warning that attackers will adapt by targeting unpatchable legacy code and human elements.

⚠️ Last week a leaked build of Anthropic's new model, Claude Capybara (also called Mythos), revealed substantially improved capabilities for automated vulnerability discovery, exploit development, and multi-step attack reasoning. The incident marks a turning point: frontier AI can compress attack lifecycles and enable scalable, novel exploitation techniques that were once the domain of advanced state actors. Security teams should treat this as a warning and accelerate risk assessments, patching, detection, and governance measures.