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 2 of 2026 · Retrospective · Covers 30 December 2025 12:00 to 6 January 2026 12:00 UTC

Read on the web

Top developments

CISA orders feds to patch MongoBleed flaw exploited in attacks

BleepingComputer · Advisory · Remediation

Zoom Stealer browser extensions harvest corporate meeting intelligence

BleepingComputer · Research · SCA

AI Model Confusion: An LLM/AI Model Supply Chain Attack

Checkmarx · Incident · SCA

Researchers Spot Modified Shai-Hulud Worm Testing Payload on npm Registry

The Hacker News · Incident · SCA

Critical AdonisJS Bodyparser Flaw (CVSS 9.2) Enables Arbitrary File Write on Servers

The Hacker News · Advisory · SCA

SAST

SastBench: A Benchmark for Testing Agentic SAST Triage

arXiv · Research · Remediation, SAST

Focus on What Matters: Fisher-Guided Adaptive Multimodal Fusion for Vulnerability Detection

arXiv · Research · SAST

Exposing Hidden Interfaces: LLM-Guided Type Inference for Reverse Engineering macOS Private Frameworks

arXiv · Research · SAST

Improving LLM-Assisted Secure Code Generation through Retrieval-Augmented-Generation and Multi-Tool Feedback

arXiv · Research · Agentic SDLC, SAST

SCA

Automated SBOM-Driven Vulnerability Triage for IoT Firmware: A Lightweight Pipeline for Risk Prioritization

arXiv · Research · Remediation, SCA

MCP-SandboxScan: WASM-based Secure Execution and Runtime Analysis for MCP Tools

arXiv · Research · AI Security, SCA

Towards a Benchmark for Dependency Decision-Making

arXiv · Research · Agentic SDLC, SCA

Supply chains, AI, and the cloud: The biggest failures (and one success) of 2025

Ars Technica: Security · Analysis · SCA

DAST

dataRLsec: Safety, Security, and Reliability With Robust Offline Reinforcement Learning for DPAs

arXiv · Research · DAST, Models

CuFuzz: Hardening CUDA Programs through Transformation and Fuzzing

arXiv · Research · DAST

Agentic SDLC

Are Copilot prompt injection flaws vulnerabilities or AI limits?

BleepingComputer · Analysis · Agentic SDLC, AI Security

VSCode IDE forks expose users to "recommended extension" attacks

BleepingComputer · Analysis · Agentic SDLC

Security in the Age of AI Teammates: An Empirical Study of Agentic Pull Requests on GitHub

arXiv · Research · Agentic SDLC

Agentic ProbLLMs: Exploiting AI Computer-Use And Coding Agents (39C3 Video + Slides)

Embrace The Red · Analysis · Agentic SDLC, AI Security

Remediation

IBM warns of critical API Connect auth bypass vulnerability

BleepingComputer · Analysis · Remediation

Advanced Vulnerability Scanning for Open Source Software: Detection and Mitigation of Log4j Vulnerabilities

arXiv · Research · Remediation

DynaFix: Iterative Automated Program Repair Driven by Execution-Level Dynamic Information

arXiv · Research · Remediation, SAST

Trust

Interoperability in AI Safety Governance: Ethics, Regulations, and Standards

arXiv · Research · Trust

Enterprise Identity Integration for AI-Assisted Developer Services: Architecture, Implementation, and Case Study

arXiv · Research · Trust

Models

Adversarial Contrastive Learning for LLM Quantization Attacks

arXiv · Research · Models

TRYLOCK: Defense-in-Depth Against LLM Jailbreaks via Layered Preference and Representation Engineering

arXiv · Research · Models

Extracting books from production language models

arXiv · Research · Models

Hidden State Poisoning Attacks against Mamba-based Language Models

arXiv · Research · Models

AI Security

Language Model Agents Under Attack: A Cross Model-Benchmark of Profit-Seeking Behaviors in Customer Service

arXiv · Research · AI Security

Agentic AI Is an Identity Problem and CISOs Will Be Accountable for the Outcome

BleepingComputer · Analysis · AI Security

Structural Representations for Cross-Attack Generalization in AI Agent Threat Detection

arXiv · Research · AI Security

OpenAI says prompt injection may never be ‘solved’ for browser agents like Atlas

CyberScoop · Analysis · AI Security

Advisories in this window

  • 106 npm advisories · 91 critical · 8 high · 7 moderate
  • 17 Packagist advisories · 11 high · 4 moderate · 2 low
  • 16 PyPI advisories · 6 high · 5 moderate · 5 low
  • 7 Go advisories · 1 critical · 1 high · 4 moderate · 1 low
  • 6 Maven advisories · 3 high · 3 moderate
  • 3 crates.io advisories · 1 critical · 1 high · 1 moderate
  • 2 Multi-ecosystem advisories · 2 moderate
  • 2 NuGet advisories · 2 moderate

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.