PREVIEW. This is how it will look in an inbox; it was not sent to you. Each real email carries its own unsubscribe link and Keryn's postal address.

Keryn Headlines

Weekly, week 8 of 2026 · Retrospective · Covers 10 February 2026 12:00 to 17 February 2026 12:00 UTC

Read on the web

Top developments

Google patches first Chrome zero-day exploited in attacks this year

BleepingComputer · Advisory · Remediation

CISA gives feds 3 days to patch actively exploited BeyondTrust flaw

BleepingComputer · Advisory · Remediation

Microsoft February 2026 Patch Tuesday fixes 6 zero-days, 58 flaws

BleepingComputer · Advisory · Remediation

CISA flags critical Microsoft SCCM flaw as exploited in attacks

BleepingComputer · Incident · Remediation

Critical BeyondTrust RCE flaw now exploited in attacks, patch now

BleepingComputer · Incident · Remediation

SAST

AXE: Grey-Box Exploitability Confirmation for Localized Vulnerability Reports

arXiv · Research · SAST

VulReaD: Knowledge-Graph-guided Software Vulnerability Reasoning and Detection

arXiv · Research · SAST

SecCodePRM: A Process Reward Model for Code Security

arXiv · Research · SAST

QRS: A Rule-Synthesizing Neuro-Symbolic Triad for Autonomous Vulnerability Discovery

arXiv · Research · Agentic SDLC, Remediation, SAST

SCA

Lazarus Campaign Plants Malicious Packages in npm and PyPI Ecosystems

The Hacker News · Incident · SCA

Weight space Detection of Backdoors in LoRA Adapters

arXiv · Research · SCA

Overthinking Loops in Agents: A Structural Risk via MCP Tools

arXiv · Research · AI Security, SCA

SkillJect: Effectively Automating Skill-Based Prompt Injection for Skill-Enabled Agents

arXiv · Research · AI Security, SCA

DAST

Following Dragons: Code Review-Guided Fuzzing

arXiv · Research · DAST

GPU-Fuzz: Finding Memory Errors in Deep Learning Frameworks

arXiv · Research · DAST

Agentic SDLC

An end-to-end agentic pipeline for smart contract translation and quality evaluation

arXiv · Research · Agentic SDLC

Assessing Spear-Phishing Website Generation in Large Language Model Coding Agents

arXiv · Research · Agentic SDLC

Agentic Knowledge Distillation: Autonomous Training of Small Language Models for SMS Threat Detection

arXiv · Research · Agentic SDLC

SWE-MiniSandbox: Container-Free Reinforcement Learning for Building Software Engineering Agents

arXiv · Research · Agentic SDLC

Remediation

Microsoft Patch Tuesday matches last year’s zero-day high with six actively exploited vulnerabilities

CyberScoop · Advisory · Remediation

Apple fixes zero-day flaw used in 'extremely sophisticated' attacks

BleepingComputer · Analysis · Remediation

Over 60 Software Vendors Issue Security Fixes Across OS, Cloud, and Network Platforms

The Hacker News · Analysis · Remediation

Windows 11 KB5077181 fixes boot failures linked to failed updates

BleepingComputer · Analysis · Remediation

Trust

Passwords to passkeys: Staying ISO 27001 compliant in a passwordless era

BleepingComputer · Standard / policy · Trust

Why ‘secure-by-design’ systems are non-negotiable in the AI era

CyberScoop · Analysis · Trust

Building the ethical AI framework of the future: from philosophy to practice

arXiv · Research · Trust

ForesightSafety Bench: A Frontier Risk Evaluation and Governance Framework towards Safe AI

arXiv · Research · Models, Trust

Models

Google says hackers are abusing Gemini AI for all attacks stages

BleepingComputer · Research · Models

Exploiting Layer-Specific Vulnerabilities to Backdoor Attack in Federated Learning

arXiv · Research · Models

Boundary Point Jailbreaking of Black-Box LLMs

arXiv · Research · Models

AI Safety Evaluations Need To Consider Cascading Effects

arXiv · Research · Models

AI Security

Manipulating AI memory for profit: The rise of AI Recommendation Poisoning

Microsoft · Research · AI Security

MCPShield: A Security Cognition Layer for Adaptive Trust Calibration in Model Context Protocol Agents

arXiv · Research · AI Security

When Benchmarks Lie: Evaluating Malicious Prompt Classifiers Under True Distribution Shift

arXiv · Research · AI Security, Models

AlignSentinel: Alignment-Aware Detection of Prompt Injection Attacks

arXiv · Research · AI Security

Advisories in this window

  • 56 npm advisories · 42 critical · 2 high · 11 moderate · 1 low
  • 24 Go advisories · 1 critical · 7 high · 13 moderate · 3 low
  • 9 PyPI advisories · 1 critical · 4 high · 2 moderate · 2 low
  • 9 crates.io advisories · 4 critical · 2 high · 2 moderate · 1 low
  • 7 Packagist advisories · 1 critical · 2 high · 4 moderate
  • 7 Maven advisories · 1 critical · 2 high · 3 moderate · 1 low
  • 1 NuGet advisory · 1 high

All advisories

Titles, links and short excerpts only; every item links to its source. Selection is automatic; the editor's note is AI-drafted and checked as described. Preview: the unsubscribe link and Keryn's postal address are added to each email when it is sent.