Skip to content
Thursday, 19 November 2026
Tenesys AI News
Subscribe

All News

TENESYS AI NEWS tracks the most important AI news and explains what changed, why it matters and whether a technology is worth testing.

“Bezpieczeństwo” — 76 articles ✕

Threat actors use agentic AI to automate credential theft and cloud exploitation

Malicious actors are increasingly deploying agentic AI frameworks, such as ARTEX, to scan public GitHub repositories for leaked credentials and immediately launch follow-on attacks against cloud services. These agents autonomously analyze output, adapt API requests upon failure, and can move from initial reconnaissance to data exfiltration from services like Azure Key Vault in minutes.

Microsoft Security
Cyber Security· Important

Cloudflare proposes new standards for AI Agent security and identity

Cloudflare is addressing the challenges posed by remotely hosted AI Agents, which obscure user identity and create accountability gaps for website operators. Because agents act as intermediaries between users and sites, traditional security mechanisms like CAPTCHAs and device fingerprinting are becoming increasingly ineffective.

Cloudflare Security
Agents· Important· 🧪 Worth Testing

Microsoft unveils PDD Kit to verify AI-generated code

Microsoft Research has introduced PDD Kit, a toolkit designed to bridge the gap between human intent and agent-generated code through Proof-Driven Development. By using formal verification via the Verus language, the kit allows developers to mathematically prove that AI-written code meets specified requirements, moving beyond traditional testing.

Microsoft Research
Models· Important· 🧪 Worth Testing

OpenAI launches GPT-6-Cyber for vulnerability analysis

OpenAI has introduced GPT-6-Cyber, a new model variant tailored for identifying, reproducing, and patching security vulnerabilities. The release requires configuration of the Daybreak environment prior to implementation.

OpenAI API Changelog
Agents· Important· 🧪 Worth Testing

Amazon Quick introduces Smart Approvals for AI agent workflows

Amazon Quick has launched Smart Approvals to automate routine agent tasks while maintaining human oversight for critical decisions. The system uses a classifier to evaluate requested actions against conversation context and user policy, preventing unnecessary interruptions without bypassing established security controls.

AWS Machine Learning
Cyber Security· Important· 🧪 Worth Testing

Cloudflare adds granular control and visibility for personal AI agents

Cloudflare has launched a new category in its Security Analytics for 'Personal Assistants,' enabling site owners to track and manage automated traffic originating from AI personal agents. Companies can now set specific access policies—allowing or blocking these agents on a per-agent basis—to protect ad revenue or sensitive areas while still facilitating legitimate user requests.

Cloudflare Security

Anthropic Introduces Watermarking for Latest Claude Models

Anthropic has updated its platform to include a text watermark for content generated by Claude Mythos 5, Opus 4.7, and Sonnet 4.6. This measure serves to identify AI-generated text, though it may be bypassed under specific technical configurations.

Claude Platform Release Notes
Cyber Security· Important· 🧪 Worth Testing

Collinear AI Releases CWE-bench v2 for Cybersecurity Evaluation

Collinear AI has launched CWE-bench v2, a benchmarking suite featuring 100 tasks covering 91 common vulnerability types. The tool allows developers to programmatically evaluate how effectively their AI agents find and fix security weaknesses in open-source repositories.

Hugging Face
Research· Important· 🧪 Worth Testing

Google releases benchmark for engineering AI agents

Google has introduced Physical-Engineering's Last Exam (PLE), a new benchmark designed to test AI agents' capabilities in mechanical, electrical, civil, and structural engineering. Unlike standard textbook tests, PLE requires agents to process large project folders, including CAD drawings and building codes, to produce compliant, safe deliverables.

Google Cloud AI
Agents· Important· 🧪 Worth Testing

Amazon Bedrock AgentCore simplifies identity propagation for enterprise agents

Amazon has updated the AgentCore Gateway to support automated token exchange between Microsoft Entra ID and protected REST APIs. By leveraging this flow, AI agents can now perform actions in platforms like ServiceNow using the identity and permissions of the end-user rather than acting as a generic service account.

AWS Machine Learning
Cyber Security· Important

NCSC assessment: The impact of AI on cyber threats through 2028

The UK's NCSC has released an assessment projecting the trajectory of AI-enabled cyber threats over the next two years. The report warns that AI will likely increase the speed and scale of cyber attacks, facilitating the rapid identification and exploitation of vulnerabilities. Three scenarios explore potential outcomes, ranging from increased global threat levels to potential mitigation through robust defensive measures.

NCSC UK
Cyber Security· Important

Data Analysis Reveals Why Most Vulnerabilities Remain Unexploited

Microsoft's analysis of 383,000 vulnerabilities since 1999 confirms that less than 0.5% are exploited in the wild. Attackers overwhelmingly favor 'takeover' vulnerabilities, such as Remote Code Execution (RCE) and privilege escalation, while AI-assisted exploit development has significantly reduced the time between disclosure and weaponization.

Microsoft Security
Cyber Security· Important· 🧪 Worth Testing

New 'CoFiBreak' Framework Enhances LLM Jailbreak Diversity

Researchers introduced CoFiBreak, a reward mechanism for training attacker LLMs that forces a higher diversity of jailbreak strategies. By using a coarse-to-fine reward structure in Group Relative Policy Optimization, the method avoids the typical reinforcement learning collapse where models converge on a narrow set of repetitive prompts.

arXiv cs.CL
Cyber Security· Important· 🧪 Worth Testing

Improved Security Monitoring for Tool-Using AI Agents

Recent research demonstrates that current security monitors often fail by confusing injection exposure with an agent's actual commitment to perform an unsafe action. The author proposes a double-stratified evaluation method that separates exposure from behavior, significantly improving the ability to predict unsafe tool calls up to two steps in advance.

arXiv cs.CL
Cyber Security· Important

Critical GitLab Integer Overflow Flaw Under Active Attack

CISA has updated its Known Exploited Vulnerabilities catalog to include CVE-2026-97229, an integer overflow issue affecting both GitLab Community and Enterprise editions. This flaw is currently being leveraged by malicious actors in real-world attacks.

CISA Advisories