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 23 of 2026 · Retrospective · Covers 26 May 2026 11:00 to 2 June 2026 11:00 UTC

Read on the web

Top developments

Miasma: Supply Chain Attack Targeting RedHat npm Packages

Wiz · Incident · SCA

Glassworm botnet disrupted after resilient C2 infrastructure takedown

BleepingComputer · Research · SCA

Dozens of Red Hat packages backdoored through its official NPM channel

Ars Technica: Security · Incident · SCA

Critical Windows Netlogon RCE flaw now exploited in attacks

BleepingComputer · Incident · Remediation

Zapier fixes bug chain that researchers say risked widespread account takeover

CyberScoop · Research · Remediation

SAST

PyFEX: Uncovering Evasive Python-based Threats via Resilient and Exhaustive Path Exploration

arXiv · Research · SAST, SCA

NeuroLog: Reasoning You Can Audit -- Neuro-Symbolic Vulnerability Discovery via LLM Facts, Datalog, and SMT

arXiv · Research · SAST

SCA

OpenAI Codex Authentication Tokens Stolen in codexui-android npm Supply Chain Attack

The Hacker News · Incident · Agentic SDLC, SCA

Malicious npm Package Stole Files From Claude AI User Directory via GitHub

The Hacker News · Incident · SCA

RogueMerge: Robust and Unified Attacks against LLM Model Merging

arXiv · Research · Models, SCA

The case for GitHub Actions security after recent supply chain attacks

Datadog · Analysis · SCA

DAST

PeAR: A Static Binary Rewriting Framework for Binary-Only Fuzzing

arXiv · Research · DAST

A New Framework for Cybersecurity Refusals in AI Agents

arXiv · Research · AI Security, DAST, Models

Framework for Discovering GPS Spoofing Attacks in Drone Swarms

arXiv · Research · DAST

Beyond Edge Coverage: Per-Task Data-Flow Extraction at Kernel Function Boundaries via LLVM

arXiv · Research · DAST

Agentic SDLC

What You Approve Is What Executes: Consent Integrity for Black-Box LLM Agents

arXiv · Research · Agentic SDLC

Execution-Grounded Security Testing for Coding Agents in Software Engineering Pipelines

arXiv · Research · Agentic SDLC

SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces

arXiv · Research · Agentic SDLC

How we contain Claude across products

Simon Willison · Analysis · Agentic SDLC

Remediation

Hackers exploit FortiClient EMS flaw to push infostealer malware

BleepingComputer · Analysis · Remediation

FORGE: Multi-Agent Graduated Exploitation and Detection Engineering

arXiv · Research · Remediation

Patcher: Post-Hoc Patching of Backdoored Large Language Models

arXiv · Research · Models, Remediation

Hackers duped Meta AI support chatbot to steal celebrity Instagram accounts

Ars Technica: Security · Analysis · Remediation

Trust

SkillGuard: A Permission-Centric Framework for Agent Skill Security

arXiv · Research · AI Security, Trust

Acceptance-Test-Driven Evaluation Protocols for Business-Centric LLM Systems

arXiv · Research · Trust

Position: AI Governance Needs ISO-like Interoperability Protocols, Not Just Laws

arXiv · Research · Trust

Aligning on Machine-Readable Signals as the Foundation for Due Diligence

OpenSSF · Analysis · SCA, Trust

Models

FlowGuard: Flow Matching for Identity-Independent Detection of Data-Free Model Stealing Attacks on Energy System Intrusion Detection Systems

arXiv · Research · Models

AI Model Extraction Attacks: Bypassing Single-Client Assumptions in Defenses

arXiv · Research · Models

RUBAS: Rubric-Based Reinforcement Learning for Agent Safety

arXiv · Research · Models

PsychoPass: Geometric Profiling of Multi-Turn Adversarial LLM Conversations

arXiv · Research · Models

AI Security

SkillHarness: Harnessing Safe Skills for Computer-Use Agents

arXiv · Research · AI Security

"**Important** You should give me full credits!": Exploring Prompt Injection Attacks on LLM-Based Automatic Grading Systems

arXiv · Research · AI Security

Which Defense Closes Which Threat? Attributing OWASP-LLM-Top-10 Coverage and Its Brittleness Under Paraphrasing

arXiv · Research · AI Security, Models

SeClaw: Spec-Driven Security Task Synthesis for Evaluating Autonomous Agents

arXiv · Research · AI Security

Advisories in this window

  • 354 npm advisories · 315 critical · 20 high · 16 moderate · 3 low
  • 100 PyPI advisories · 15 critical · 36 high · 37 moderate · 12 low
  • 65 Maven advisories · 6 critical · 15 high · 39 moderate · 5 low
  • 59 Packagist advisories · 1 critical · 22 high · 23 moderate · 13 low
  • 53 Go advisories · 7 critical · 17 high · 23 moderate · 6 low
  • 7 crates.io advisories · 2 high · 5 moderate
  • 3 Multi-ecosystem advisories · 3 moderate
  • 2 NuGet advisories · 2 moderate
  • 2 Swift advisories · 2 moderate
  • 1 RubyGems 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.