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 36 of 2025 · Retrospective · Covers 26 August 2025 11:00 to 2 September 2025 11:00 UTC

Read on the web

Top developments

WhatsApp patches vulnerability exploited in zero-day attacks

BleepingComputer · Advisory · Remediation

s1ngularity: supply chain attack leaks secrets on GitHub: everything you need to know

Wiz · Incident · SCA

Malware devs abuse Anthropic’s Claude AI to build ransomware

BleepingComputer · Incident · Agentic SDLC

Researchers flag code that uses AI systems to carry out ransomware attacks

CyberScoop · Incident · AI Security

Hundreds of Salesforce customers impacted by attack spree linked to third-party AI agent

CyberScoop · Research · AI Security

SAST

VULSOLVER: Vulnerability Detection via LLM-Driven Constraint Solving

arXiv · Research · SAST

Detecting Stealthy Data Poisoning Attacks in AI Code Generators

arXiv · Research · Models, SAST

AVIATOR: Towards AI-Agentic Vulnerability Injection Workflow for High-Fidelity, Large-Scale Code Security Dataset

arXiv · Research · AI Security, SAST

SIExVulTS: Sensitive Information Exposure Vulnerability Detection System using Transformer Models and Static Analysis

arXiv · Research · SAST

SCA

Malicious npm Package nodejs-smtp Mimics Nodemailer, Targets Atomic and Exodus Wallets

The Hacker News · Research · SCA

Policy-driven Software Bill of Materials on GitHub: An Empirical Study

arXiv · Research · SCA

An Empirical Study of Vulnerable Package Dependencies in LLM Repositories

arXiv · Research · SCA

Researchers Find VS Code Flaw Allowing Attackers to Republish Deleted Extensions Under Same Names

The Hacker News · Research · SCA

DAST

WFC/WFD: Web Fuzzing Commons, Dataset and Guidelines to Support Experimentation in REST API Fuzzing

arXiv · Research · DAST

Locus: Agentic Predicate Synthesis for Directed Fuzzing

arXiv · Research · DAST

Multi-Agent Penetration Testing AI for the Web

arXiv · Research · Agentic SDLC, DAST

Infosec products of the month: August 2025

Help Net Security · Analysis · DAST

Agentic SDLC

AWS Kiro: Arbitrary Code Execution via Indirect Prompt Injection

Embrace The Red · Analysis · Agentic SDLC, AI Security

Wrap Up: The Month of AI Bugs

Embrace The Red · Analysis · Agentic SDLC, AI Security

AgentHopper: An AI Virus

Embrace The Red · Analysis · Agentic SDLC, AI Security

Human-Written vs. AI-Generated Code: A Large-Scale Study of Defects, Vulnerabilities, and Complexity

arXiv · Research · Agentic SDLC

Remediation

Passwordstate dev urges users to patch auth bypass vulnerability

BleepingComputer · Analysis · Remediation

Microsoft says recent Windows update didn't kill your SSD

BleepingComputer · Analysis · Remediation

How we accelerated Secret Protection engineering with Copilot

GitHub · Analysis · Agentic SDLC, Remediation

Trust

Poison Once, Refuse Forever: Weaponizing Alignment for Injecting Bias in LLMs

arXiv · Research · Models, Trust

Private, Verifiable, and Auditable AI Systems

arXiv · Research · Trust

Releases now support immutability in public preview

GitHub · Analysis · SCA, Trust

A NIS2 pan-European registry for identifying and classifying essential and important entities

arXiv · Research · Trust

Models

Enhancing Reliability in LLM-Integrated Robotic Systems: A Unified Approach to Security and Safety

arXiv · Research · AI Security, Models

Augmented Shuffle Differential Privacy Protocols for Large-Domain Categorical and Key-Value Data

arXiv · Research · Models

Partially Functional Dynamic Backdoor Diffusion-based Causal Model

arXiv · Research · Models

The Resurgence of GCG Adversarial Attacks on Large Language Models

arXiv · Research · Models

AI Security

Cybersecurity AI: Hacking the AI Hackers via Prompt Injection

arXiv · Research · AI Security

A Whole New World: Creating a Parallel-Poisoned Web Only AI-Agents Can See

arXiv · Research · AI Security

Windsurf MCP Integration: Missing Security Controls Put Users at Risk

Embrace The Red · Analysis · Agentic SDLC, AI Security

PromptSleuth: Detecting Prompt Injection via Semantic Intent Invariance

arXiv · Research · AI Security

Advisories in this window

  • 25 PyPI advisories · 9 high · 14 moderate · 2 low
  • 16 Go advisories · 1 critical · 8 high · 6 moderate · 1 low
  • 11 npm advisories · 2 critical · 4 high · 5 moderate
  • 6 Packagist advisories · 1 critical · 1 high · 4 moderate
  • 4 Maven advisories · 1 critical · 2 moderate · 1 low
  • 4 NuGet advisories · 2 high · 1 moderate · 1 low
  • 3 crates.io advisories · 2 moderate · 1 low
  • 2 RubyGems advisories · 2 moderate
  • 1 GitHub Actions 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.