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 50 of 2025 · Retrospective · Covers 2 December 2025 12:00 to 9 December 2025 12:00 UTC

Read on the web

Top developments

Microsoft Silently Patches Windows LNK Flaw After Years of Active Exploitation

The Hacker News · Advisory · Remediation

Taking Down More Malicious VSCode Extensions in the ‘Prettier’ Campaign

Checkmarx · Incident · SCA

ThreatsDay Bulletin: Wi-Fi Hack, npm Worm, DeFi Theft, Phishing Blasts— and 15 More Stories

The Hacker News · Incident · SCA

Researchers Find Malicious VS Code, Go, npm, and Rust Packages Stealing Developer Data

The Hacker News · Research · Agentic SDLC, SCA

Researcher Uncovers 30+ Flaws in AI Coding Tools Enabling Data Theft and RCE Attacks

The Hacker News · Research · AI Security

SAST

USCSA: Evolution-Aware Security Analysis for Proxy-Based Upgradeable Smart Contracts

arXiv · Research · SAST

VulnLLM-R: Specialized Reasoning LLM with Agent Scaffold for Vulnerability Detection

arXiv · Research · SAST

SCA

React2Shell (CVE-2025-55182): Everything You Need to Know About the Critical React Vulnerability

Wiz · Analysis · DAST, Remediation, SCA

Secure or Suspect? Investigating Package Hallucinations of Shell Command in Original and Quantized LLMs

arXiv · Research · SCA

Malicious VSCode extensions on Microsoft's registry drop infostealers

BleepingComputer · Analysis · SCA

Automated Code Review Assignments: An Alternative Perspective of Code Ownership on GitHub

arXiv · Research · SCA

DAST

Beyond Model Jailbreak: Systematic Dissection of the "Ten DeadlySins" in Embodied Intelligence

arXiv · Research · DAST, Models

Bootstrapping Fuzzers for Compilers of Low-Resource Language Dialects Using Language Models

arXiv · Research · DAST

BGPFuzz: Automated Configuration Fuzzing of the Border Gateway Protocol

arXiv · Research · DAST

Targeted Testing of Compiler Optimizations via Grammar-Level Composition Styles

arXiv · Research · DAST

Agentic SDLC

DeepCode: Open Agentic Coding

arXiv · Research · Agentic SDLC

WildCode Revisited: A Comprehensive Empirical Study on the Security of LLM-Generated Code

arXiv · Research · Agentic SDLC

Is Vibe Coding Safe? Benchmarking Vulnerability of Agent-Generated Code in Real-World Tasks

arXiv · Research · Agentic SDLC, SCA

Exploiting Markdown Injection in AI agents: Microsoft Copilot Chat and Google Gemini

Checkmarx · Analysis · Agentic SDLC, AI Security

Remediation

PBFuzz: Agentic Directed Fuzzing for PoV Generation

arXiv · Research · Remediation

Trust

AgentCrypt: Advancing Privacy and (Secure) Computation in AI Agent Collaboration

arXiv · Research · AI Security, Trust

A Technical Policy Blueprint for Trustworthy Decentralized AI

arXiv · Research · Trust

Recap: OpenSSF Community Day Korea 2025

OpenSSF · Analysis · Trust

The Decision Path to Control AI Risks Completely: Fundamental Control Mechanisms for AI Governance

arXiv · Research · Trust

Models

Optimal Perturbation Budget Allocation for Data Poisoning in Offline Reinforcement Learning

arXiv · Research · Models

Fully Decentralized Certified Unlearning

arXiv · Research · Models

A Practical Framework for Evaluating Medical AI Security: Reproducible Assessment of Jailbreaking and Privacy Vulnerabilities Across Clinical Specialties

arXiv · Research · Models

Improving the Sensitivity of Backdoor Detectors via Class Subspace Orthogonalization

arXiv · Research · Models

AI Security

Architecting Security for Agentic Capabilities in Chrome

Google · Analysis · AI Security

Attention is All You Need to Defend Against Indirect Prompt Injection Attacks in LLMs

arXiv · Research · AI Security

Systematization of Knowledge: Security and Safety in the Model Context Protocol Ecosystem

arXiv · Research · AI Security

UK cyber agency warns LLMs will always be vulnerable to prompt injection

CyberScoop · Analysis · AI Security

Advisories in this window

  • 57 npm advisories · 50 critical · 4 high · 2 moderate · 1 low
  • 28 Go advisories · 3 critical · 11 high · 12 moderate · 2 low
  • 13 PyPI advisories · 2 critical · 4 high · 5 moderate · 2 low
  • 8 Packagist advisories · 1 critical · 3 high · 3 moderate · 1 low
  • 7 Maven advisories · 1 critical · 3 high · 2 moderate · 1 low
  • 5 crates.io advisories · 2 high · 2 moderate · 1 low
  • 2 RubyGems advisories · 2 critical
  • 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.