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 20 of 2025 · Retrospective · Covers 6 May 2025 11:00 to 13 May 2025 11:00 UTC

Read on the web

Top developments

SonicWall customers confront resurgence of actively exploited vulnerabilities

CyberScoop · Incident · Trust

Critical Langflow RCE flaw exploited to hack AI app servers

BleepingComputer · Incident · Remediation

ASUS DriverHub flaw let malicious sites run commands with admin rights

BleepingComputer · Analysis · Remediation

Malicious PyPi package hides RAT malware, targets Discord devs since 2022

BleepingComputer · Analysis · SCA

SysOwned, Your Friendly Support Ticket - SysAid On-Premise Pre-Auth RCE Chain (CVE-2025-2775 And Friends)

watchTowr · Analysis · Remediation

SAST

AutoPatch: Multi-Agent Framework for Patching Real-World CVE Vulnerabilities

arXiv · Research · Remediation, SAST

SCA

Malicious npm Packages Infect 3,200+ Cursor Users With Backdoor, Steal Credentials

The Hacker News · Research · SCA

Wide World of Cyber: How state adversaries attack security vendors

Risky Business · Analysis · SCA

DAST

RedTeamLLM: an Agentic AI framework for offensive security

arXiv · Research · DAST

Offensive Security for AI Systems: Concepts, Practices, and Applications

arXiv · Research · DAST

An Empirical Study of Fuzz Harness Degradation

arXiv · Research · DAST

Directed Greybox Fuzzing via Large Language Model

arXiv · Research · DAST

Agentic SDLC

Ox Security lands a fresh $60M to scan for vulnerabilities in code

TechCrunch: Security · Analysis · Agentic SDLC

Remediation

Cisco fixes max severity IOS XE flaw letting attackers hijack devices

BleepingComputer · Incident · Remediation

SonicWall urges admins to patch VPN flaw exploited in attacks

BleepingComputer · Advisory · Remediation

Microsoft: April updates cause Windows Server auth issues

BleepingComputer · Analysis · Remediation

Beyond Vulnerability Management – Can You CVE What I CVE?

The Hacker News · Analysis · Remediation

Trust

sigstore-go 1.0 is now available

Sigstore · Release · Trust

Getting Ready for the EU AI Act in Healthcare. A call for Sustainable AI Development and Deployment

arXiv · Research · Trust

OpenSSF Tech Talk Recap: Using the OSPS Baseline to Navigate Standards and Regulations

OpenSSF · Analysis · Trust

From Glue-Code to Protocols: A Critical Analysis of A2A and MCP Integration for Scalable Agent Systems

arXiv · Research · AI Security, Trust

Models

MixBridge: Heterogeneous Image-to-Image Backdoor Attack through Mixture of Schrödinger Bridges

arXiv · Research · Models

Decoding One Safety Trigger Token for Balancing Safety and Usability in Large Language Models

arXiv · Research · Models

TokenProber: Jailbreaking Text-to-image Models via Fine-grained Word Impact Analysis

arXiv · Research · Models

Practical Reasoning Interruption Attacks on Reasoning Large Language Models

arXiv · Research · AI Security, Models

AI Security

System Prompt Poisoning: Persistent Attacks on Large Language Models Beyond User Injection

arXiv · Research · AI Security, Models

AgentVigil: Generic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents

arXiv · Research · AI Security

Threat Modeling for AI: The Case for an Asset-Centric Approach

arXiv · Research · AI Security

Defending against Indirect Prompt Injection by Instruction Detection

arXiv · Research · AI Security

Advisories in this window

  • 81 npm advisories · 76 critical · 2 high · 2 moderate · 1 low
  • 9 Maven advisories · 5 high · 3 moderate · 1 low
  • 8 crates.io advisories · 3 moderate · 5 low
  • 3 PyPI advisories · 1 high · 1 moderate · 1 low
  • 3 RubyGems advisories · 1 high · 2 moderate
  • 3 Packagist advisories · 3 moderate
  • 3 Go advisories · 1 critical · 1 high · 1 low
  • 1 GitHub Actions advisory · 1 moderate
  • 1 Multi-ecosystem advisory · 1 moderate
  • 1 NuGet advisory · 1 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.