Articles by tag: ai-security

Intel hardware security competition 2026: problems and vulnerabilities

Intel announced the Hardware Security Award competition without a cash prize amid 13 vulnerabilities. We analyze hidden motives, Resilient Silicon issues, and forecasts until August 2026. Read the analysis.

MCP Vulnerability from Anthropic: RCE and Double Standards

Critical RCE vulnerability in the MCP protocol from Anthropic is ignored by the company despite responsible disclosure. Analysis of attack vectors and recommendations for developers.

Open AI Models Found Vulnerabilities Like Mythos

Researchers Reproduced Results of the Closed Mythos Model Using GPT-5.4 and Claude Opus. Detailed Breakdown of Technical Details and Conclusions.

AI Security in Jira: Risks and Recommendations for DevSecOps

How to Protect Data When Integrating AI Agents into Jira? We Analyze Preprocessor Risks, the Danger of Personal Accounts, and Provide Practical Recommendations for DevSecOps Engineers.

Hybrid RAG without cloud: search in corporate documents

How to implement hybrid RAG for document search without data leaks? We break down the architecture of a self-hosted solution for regulated industries. Learn more.

LLM Security: Evolution from Firewall to GAF | In-Depth Analysis

How modern threats to LLM require multi-level protection. We analyze incidents of 2025-2026 and the transition to Generative Application Firewall. Learn how to protect AI systems.

RCE in Marimo CVE-2026-39987: exploit in 9 hours

Critical pre-auth RCE in Marimo up to 0.23.0 gives root-shell via WebSocket. Sysdig: attack 9 hours after advisory. Update, rotate secrets, protect AI toolchain. Detailed analysis and measures.

McKinsey AI Vulnerability: Access to Data in Hours

Critical hole in Lilli system allowed extracting millions of documents. Risk analysis for business and AI platforms. Learn how to protect corporate systems from SQL injections.

Claude Mythos found a bug in OpenBSD 27 years

Claude Mythos Preview detected vulnerabilities in OpenBSD, FFmpeg, Linux. Benchmarks, partners, funding. Learn how AI changes security audits for developers.

AI Giants' Collaboration Against Distillation

OpenAI, Anthropic and Google are sharing data on attacks by DeepSeek and Chinese labs. Details on adversarial distillation, scales, and political context. For AI developers.

Agentic SAMM for secure agents

Learn how ASAMM extends OWASP SAMM for agentic systems: threat taxonomy, 17 controls, trust model. Integrate into SDLC for middle/senior devs. Study the framework now.

AI Traps in B2B: Shadow AI and Implementation Failures

Analysis of three B2B AI implementation errors: from leaks through personal accounts to overpayments for agents. Steps for control, templates, and adaptation. Learn how to avoid losses of millions.

Agentic SOC in 2026: AI agents for SOC

Breaking down Agentic SOC: advantages of AI agents in triage and response, architecture with guardrails, risks and implementation. For IB specialists — checklists and stages without illusions.