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 9 of 2026 · Retrospective · Covers 17 February 2026 12:00 to 24 February 2026 12:00 UTC

Read on the web

Top developments

CISA: Recently patched RoundCube flaws now exploited in attacks

BleepingComputer · Advisory · Remediation

CISA orders feds to patch actively exploited Dell flaw within 3 days

BleepingComputer · Advisory · Remediation

Cline CLI 2.3.0 Supply Chain Attack Installed OpenClaw on Developer Systems

The Hacker News · Incident · SCA

ThreatsDay Bulletin: OpenSSL RCE, Foxit 0-Days, Copilot Leak, AI Password Flaws & 20+ Stories

The Hacker News · Incident · Agentic SDLC, AI Security

Malicious npm Packages Harvest Crypto Keys, CI Secrets, and API Tokens

The Hacker News · Research · SCA

SAST

Automatic, Expressive, and Scalable Fuzzing with Stitching

arXiv · Research · DAST, SAST

SCA

SOK: A Taxonomy of Attack Vectors and Defense Strategies for Agentic Supply Chain Runtime

arXiv · Research · AI Security, SCA

On the Variability of Source Code in Maven Package Rebuilds

arXiv · Research · Agentic SDLC, SCA

Trustworthy AI LLM Scalability Risk Index (LSRI): A Cybersecurity Framework Assessing Agentic-AI Security & Software Model Supply Chain Safety Boosting AI-Generated Malware Defense & Explainability Mitigating Emerging Risks of Generative AI

arXiv · Research · AI Security, DAST, Models, SCA

Package Managers à la Carte: A Formal Model of Dependency Resolution

arXiv · Research · SCA

DAST

ICSSPulse: A Modular LLM-Assisted Platform for Industrial Control System Penetration Testing

arXiv · Research · DAST

FuzzySQL: Uncovering Hidden Vulnerabilities in DBMS Special Features with LLM-Driven Fuzzing

arXiv · Research · DAST

Can AI Lower the Barrier to Cybersecurity? A Human-Centered Mixed-Methods Study of Novice CTF Learning

arXiv · Research · DAST

What Makes a Good LLM Agent for Real-world Penetration Testing?

arXiv · Research · DAST

Agentic SDLC

AdapTools: Adaptive Tool-based Indirect Prompt Injection Attacks on Agentic LLMs

arXiv · Research · Agentic SDLC, AI Security

When AI Teammates Meet Code Review: Collaboration Signals Shaping the Integration of Agent-Authored Pull Requests

arXiv · Research · Agentic SDLC

GraphSkill: Documentation-Guided Hierarchical Retrieval-Augmented Coding for Complex Graph Reasoning

arXiv · Research · Agentic SDLC

Can Adversarial Code Comments Fool AI Security Reviewers -- Large-Scale Empirical Study of Comment-Based Attacks and Defenses Against LLM Code Analysis

arXiv · Research · Agentic SDLC, Models

Remediation

Red-Teaming Claude Opus and ChatGPT-based Security Advisors for Trusted Execution Environments

arXiv · Research · Models, Remediation

Microsoft Patches CVE-2026-26119 Privilege Escalation in Windows Admin Center

The Hacker News · Analysis · Remediation

AI Found Twelve New Vulnerabilities in OpenSSL

Schneier on Security · Analysis · Remediation

Anthropic Launches Claude Code Security for AI-Powered Vulnerability Scanning

The Hacker News · Announcement · Agentic SDLC, Remediation

Trust

EU Cyber Resilience Act (CRA) in Practice @ FOSDEM 2026: From Awareness to Action

OpenSSF · Standard / policy · Trust

Right to History: A Sovereignty Kernel for Verifiable AI Agent Execution

arXiv · Research · AI Security, Trust

OpenPort Protocol: A Security Governance Specification for AI Agent Tool Access

arXiv · Research · AI Security, Trust

From Docs to Descriptions: Smell-Aware Evaluation of MCP Server Descriptions

arXiv · Research · Trust

Models

CREDIT: Certified Ownership Verification of Deep Neural Networks Against Model Extraction Attacks

arXiv · Research · Models

CITED: A Decision Boundary-Aware Signature for GNNs Towards Model Extraction Defense

arXiv · Research · Models

CIBER: A Comprehensive Benchmark for Security Evaluation of Code Interpreter Agents

arXiv · Research · AI Security, Models

When Backdoors Go Beyond Triggers: Semantic Drift in Diffusion Models Under Encoder Attacks

arXiv · Research · Models

AI Security

ICON: Indirect Prompt Injection Defense for Agents based on Inference-Time Correction

arXiv · Research · AI Security

Skill-Inject: Measuring Agent Vulnerability to Skill File Attacks

arXiv · Research · AI Security

The LLMbda Calculus: AI Agents, Conversations, and Information Flow

arXiv · Research · AI Security

Security Risks of AI Agents Hiring Humans: An Empirical Marketplace Study

arXiv · Research · AI Security

Advisories in this window

  • 138 npm advisories · 34 critical · 56 high · 39 moderate · 9 low
  • 29 PyPI advisories · 4 critical · 11 high · 10 moderate · 4 low
  • 23 Go advisories · 2 critical · 7 high · 10 moderate · 4 low
  • 22 Packagist advisories · 1 critical · 7 high · 11 moderate · 3 low
  • 11 Maven advisories · 2 critical · 4 high · 2 moderate · 3 low
  • 6 crates.io advisories · 4 high · 1 moderate · 1 low
  • 2 RubyGems advisories · 1 high · 1 moderate
  • 1 Swift advisory · 1 moderate
  • 1 GitHub Actions advisory · 1 moderate
  • 1 Multi-ecosystem 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.