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 22 of 2026 · Retrospective · Covers 19 May 2026 11:00 to 26 May 2026 11:00 UTC

Read on the web

Top developments

CISA orders feds to patch actively exploited Drupal vulnerability

BleepingComputer · Incident · Remediation

Microsoft warns of new Defender zero-days exploited in attacks

BleepingComputer · Incident · Remediation

CVE-2009-1537: Microsoft DirectX NULL Byte Overwrite Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory · Agentic SDLC

GitHub confirms breach of 3,800 repos via malicious VSCode extension

BleepingComputer · Incident · SCA

Laravel Lang packages hijacked to deploy credential-stealing malware

BleepingComputer · Analysis · SCA

SAST

Heimdall: Formally Verified Automated Migration of Legacy eBPF Programs to Rust

arXiv · Research · SAST

Enhancing Reliability in LLM-Based Secure Code Generation

arXiv · Research · SAST

Automated Repair of TEE Partitioning Issues via DSL-Guided and LLM-Assisted Patching

arXiv · Research · Remediation, SAST

Quality and Security Signals in AI-Generated Python Refactoring Pull Requests

arXiv · Research · SAST

SCA

GHSA-ghwc-95x2-682j / CVE-2026-9082: Drupal Core has a SQL Injection issue

GitHub Advisory Database · Advisory · SCA

Staged publishing and new install-time controls for npm

GitHub · Analysis · SCA

A hacker group is poisoning open source code at an unprecedented scale

Ars Technica: Security · Analysis · SCA

GitHub links repo breach to TanStack npm supply-chain attack

BleepingComputer · Incident · SCA

DAST

Detecting Malicious Packages using the OSV API

OpenSSF · Research · DAST, SCA

Batch Me If You Can: Coverage-guided RPKI Fuzzing at Scale

arXiv · Research · DAST

APT-Agent: Automated Penetration Testing using Large Language Models

arXiv · Research · DAST

Are Frontier LLMs Ready for Cybersecurity? Evidence for Vertical Foundation Models from Dual-Mode Vulnerability Benchmarks

arXiv · Research · Agentic SDLC, AI Security, DAST

Agentic SDLC

SEC-bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks?

arXiv · Research · Agentic SDLC

Anthropic’s restricted Claude Mythos model may be coming to Claude Code

BleepingComputer · Analysis · Agentic SDLC

From Preventive to Reactive: How AI Coding Assistants Transform Developers' Security Awareness

arXiv · Research · Agentic SDLC

Philosophical Dispositions as Behavioral Constraints for AI-Assisted Code Review: An Empirical Study

arXiv · Research · Agentic SDLC

Remediation

Max severity Cisco Secure Workload flaw gives Site Admin privileges

BleepingComputer · Analysis · Remediation

KnowledgeDeliver LMS Flaw Exploited to Deploy Godzilla and Cobalt Strike

The Hacker News · Incident · Remediation

Microsoft: Domain Controller lookup may fail on Windows Server 2016

BleepingComputer · Analysis · Remediation

Ubiquiti patches three max severity UniFi OS vulnerabilities

BleepingComputer · Analysis · Remediation

Trust

Proof of Useful Attestation: A Consensus Primitive for Attestation-Native Chains

arXiv · Research · Trust

Who judges the judges? Governance from metrics: a runtime framework for continuous LLM compliance monitoring

arXiv · Research · Trust

OpenSSF Notes Quarter of Growth with New Members, Added AI Security Resources, and Growing Community

OpenSSF · Analysis · Trust

An Evidence-driven Protocol for Trustworthy CI Pipelines

arXiv · Research · SCA, Trust

Models

Cordyceps: Covert Control Attacks on LLMs via Data Poisoning

arXiv · Research · AI Security, Models

GradSentry: Gradient Spectral Entropy for Backdoor Sample Filtering in Large Language Model Fine-Tuning

arXiv · Research · Models

Open-Weight LLM Fine-Tuning Defenses are Susceptible to Simple Attacks

arXiv · Research · Models

Jailbreak susceptibility prediction and mitigation via the behavioral geometry of models

arXiv · Research · Models

AI Security

Cordon-MAS: Defending RAG against Knowledge Poisoning via Information-Flow Control

arXiv · Research · AI Security

Aligning Provenance with Authorization: A Dual-Graph Defense for LLM Agents

arXiv · Research · AI Security

AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents

arXiv · Research · AI Security

How Agentic AI Coding Assistants Become the Attacker's Shell

arXiv · Research · AI Security

Advisories in this window

  • 183 npm advisories · 134 critical · 23 high · 21 moderate · 5 low
  • 71 Packagist advisories · 4 critical · 21 high · 24 moderate · 22 low
  • 49 Go advisories · 7 critical · 16 high · 19 moderate · 7 low
  • 31 PyPI advisories · 5 critical · 13 high · 11 moderate · 2 low
  • 19 Maven advisories · 3 critical · 8 high · 7 moderate · 1 low
  • 12 crates.io advisories · 5 high · 7 moderate
  • 11 NuGet advisories · 1 high · 8 moderate · 2 low
  • 3 Hex advisories · 3 high
  • 2 Multi-ecosystem advisories · 2 critical
  • 2 GitHub Actions 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.