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 44 of 2025 · Retrospective · Covers 21 October 2025 11:00 to 28 October 2025 11:00 UTC

Read on the web

Top developments

Chrome Zero-Day Exploited to Deliver Italian Memento Labs' LeetAgent Spyware

The Hacker News · Incident · Remediation

Newly Patched Critical Microsoft WSUS Flaw Comes Under Active Exploitation

The Hacker News · Advisory · Remediation

TP-Link warns of critical command injection flaw in Omada gateways

BleepingComputer · Analysis · Remediation

Top security researcher shares their bug bounty process

GitHub · Research · Agentic SDLC

Self-Spreading 'GlassWorm' Infects VS Code Extensions in Widespread Supply Chain Attack

The Hacker News · Incident · SCA

SAST

Beyond Imprecise Distance Metrics: Trace-Guided Directed Greybox Fuzzing via LLM-Predicted Call Stacks

arXiv · Research · SAST

LLM-Powered Detection of Price Manipulation in DeFi

arXiv · Research · SAST

Learning to Triage Vulnerability Reports from Program Analysis: An Empirical Study in Node.js

arXiv · Research · Remediation, SAST

Collaborative penetration testing suite for emerging generative AI algorithms

arXiv · Research · Models, SAST

SCA

ThreatsDay Bulletin: $176M Crypto Fine, Hacking Formula 1, Chromium Vulns, AI Hijack & More

The Hacker News · Incident · SCA

ProGQL: A Provenance Graph Query System for Cyber Attack Investigation

arXiv · Research · SCA

Towards Socio-Technical Topology-Aware Adaptive Threat Detection in Software Supply Chains

arXiv · Research · SCA

NeuPerm: Disrupting Malware Hidden in Neural Network Parameters by Leveraging Permutation Symmetry

arXiv · Research · SCA

Agentic SDLC

QueryIPI: Query-agnostic Indirect Prompt Injection on Coding Agents

arXiv · Research · Agentic SDLC, AI Security, Models

New CoPhish attack steals OAuth tokens via Copilot Studio agents

BleepingComputer · Analysis · Agentic SDLC, AI Security

Living dangerously with Claude

Simon Willison · Analysis · Agentic SDLC

Cursor, Windsurf IDEs riddled with 94+ n-day Chromium vulnerabilities

BleepingComputer · Analysis · Agentic SDLC, Remediation

Remediation

Sharepoint ToolShell attacks targeted orgs across four continents

BleepingComputer · Analysis · Remediation

Critical WSUS flaw in Windows Server now exploited in attacks

BleepingComputer · Research · Remediation

Windows Server emergency patches fix WSUS bug with PoC exploit

BleepingComputer · Research · Remediation

QNAP warns of critical ASP.NET flaw in its Windows backup software

BleepingComputer · Analysis · Remediation

Trust

SBOMs in the Era of the CRA: Toward a Unified and Actionable Framework

OpenSSF · Standard / policy · Trust

Beyond the Checkbox: How Wiz Transforms SOC 2 into a Security Powerhouse

Wiz · Analysis · Trust

Embedding Explainable AI in NHS Clinical Safety: The Explainability-Enabled Clinical Safety Framework (ECSF)

arXiv · Research · Trust

The Nuclear Analogy in AI Governance Research

arXiv · Research · Trust

Models

When Compression Becomes an Attack Surface: Black-Box Attacks on Prompt-Compressed LLM Agents

arXiv · Research · AI Security, Models

Is Your Prompt Poisoning Code? Defect Induction Rates and Security Mitigation Strategies

arXiv · Research · Models

Sentra-Guard: A Real-Time Multilingual Defense Against Adversarial LLM Prompts

arXiv · Research · AI Security, Models

Breaking Agent Backbones: Evaluating the Security of Backbone LLMs in AI Agents

arXiv · Research · AI Security, Models

AI Security

MCPGuard : Automatically Detecting Vulnerabilities in MCP Servers

arXiv · Research · AI Security

AgentBound: Securing Execution Boundaries of AI Agents

arXiv · Research · AI Security

Soft Instruction De-escalation Defense

arXiv · Research · AI Security

GhostEI-Bench: Do Mobile Agents Resilience to Environmental Injection in Dynamic On-Device Environments?

arXiv · Research · AI Security

Advisories in this window

  • 125 npm advisories · 120 critical · 2 high · 3 moderate
  • 27 Maven advisories · 4 high · 18 moderate · 5 low
  • 15 Go advisories · 3 critical · 6 high · 6 moderate
  • 15 Packagist advisories · 2 high · 11 moderate · 2 low
  • 10 PyPI advisories · 3 high · 6 moderate · 1 low
  • 7 crates.io advisories · 2 high · 1 moderate · 4 low
  • 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.