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 19 of 2026 · Retrospective · Covers 28 April 2026 11:00 to 5 May 2026 11:00 UTC

Read on the web

Top developments

Supply Chain Campaign Targets SAP npm Packages with Credential-Stealing Malware

Wiz · Incident · SCA

CISA orders feds to patch Windows flaw exploited as zero-day

BleepingComputer · Advisory · Remediation

Backdoored PyTorch Lightning package drops credential stealer

BleepingComputer · Incident · SCA

CanisterSprawl: pgserve Compromised on npm: Malicious Versions Harvest Credentials and Exfiltrate to a Decentralized ICP Canister

StepSecurity · Incident · SCA

elementary-data Compromised on PyPI and GHCR: Forged Release Pushed via GitHub Actions Script Injection

StepSecurity · Incident · SCA

SAST

QASecClaw: A Multi-Agent LLM Approach for False Positive Reduction in Static Application Security Testing

arXiv · Research · SAST

MCPHunt: An Evaluation Framework for Cross-Boundary Data Propagation in Multi-Server MCP Agents

arXiv · Research · AI Security, SAST

VulStyle: A Multi-Modal Pre-Training for Code Stylometry-Augmented Vulnerability Detection

arXiv · Research · SAST

An Empirical Security Evaluation of LLM-Generated Cryptographic Rust Code

arXiv · Research · SAST

SCA

ScarCruft hackers push BirdCall Android malware via game platform

BleepingComputer · Analysis · SCA

Bitwarden CLI Hijacked on npm: Bun-Staged Credential Stealer Targets Developers, GitHub Actions, and AI Tools

StepSecurity · Analysis · SCA

Why a recent supply-chain attack singled out security firms Checkmarx and Bitwarden

Ars Technica: Security · Incident · SCA

Cryptographic Registry Provenance: Structural Defense Against Dependency Confusion in AI Package Ecosystems

arXiv · Research · SCA

DAST

FunFuzz: An LLM-Powered Evolutionary Fuzzing Framework

arXiv · Research · DAST

Trace: Unmasking AI Attack Agents Through Terminal Behavior Fingerprinting

arXiv · Research · AI Security, DAST

ClozeMaster: Fuzzing Rust Compiler by Harnessing LLMs for Infilling Masked Real Programs

arXiv · Research · DAST

Autonomous LLM Agents & CTFs: A Second Look

arXiv · Research · DAST

Agentic SDLC

GitHub fixes RCE flaw that gave access to millions of private repos

BleepingComputer · Analysis · Agentic SDLC, Remediation

Shai-Hulud Strikes SAP: Supply Chain Worm Weaponized Claude Code to Compromise the CAP Framework

Mend.io · Incident · Agentic SDLC, AI Security

TRE Python binding — ReDoS robustness demo

Simon Willison · Analysis · Agentic SDLC

Copirate 365 at DEF CON: Plundering in the Depths of Microsoft Copilot (CVE-2026-24299)

Embrace The Red · Analysis · Agentic SDLC, AI Security, Remediation

Remediation

cPanel, WHM emergency update fixes critical auth bypass bug

BleepingComputer · Analysis · Remediation

Progress warns of critical MOVEit Automation auth bypass flaw

BleepingComputer · Analysis · Remediation

APIOT: Autonomous Vulnerability Management Across Bare-Metal Industrial OT Networks

arXiv · Research · Remediation

CyberAId: AI-Driven Cybersecurity for Financial Service Providers

arXiv · Research · Remediation

Trust

ARGUS: Defending LLM Agents Against Context-Aware Prompt Injection

arXiv · Research · AI Security, Trust

When Agents Handle Secrets: A Survey of Confidential Computing for Agentic AI

arXiv · Research · AI Security, Trust

Human-Provenance Verification should be Treated as Labor Infrastructure in AI-Saturated Markets

arXiv · Research · Trust

Phionyx: A Deterministic AI Runtime Architecture with Structured State Management and Pre-Response Governance

arXiv · Research · Trust

Models

Self-Mined Hardness for Safety Fine-Tuning

arXiv · Research · Models

Revisiting JBShield: Breaking and Rebuilding Representation-Level Jailbreak Defenses

arXiv · Research · Models

MITRE ATLAS (releases) v5.6.0

MITRE · Release · AI Security, Models

ContextualJailbreak: Evolutionary Red-Teaming via Simulated Conversational Priming

arXiv · Research · Models

AI Security

Learning from the Vercel breach: Shadow AI & OAuth sprawl

BleepingComputer · Incident · AI Security

MEMSAD: Gradient-Coupled Anomaly Detection for Memory Poisoning in Retrieval-Augmented Agents

arXiv · Research · AI Security

PIIGuard: Mitigating PII Harvesting under Adversarial Sanitization

arXiv · Research · AI Security

Trojan Hippo Bench: A Dynamic Benchmark for Persistent Memory Attacks and Defenses in LLM Agents

arXiv · Research · AI Security

Advisories in this window

  • 130 npm advisories · 95 critical · 13 high · 19 moderate · 3 low
  • 48 Go advisories · 4 critical · 25 high · 16 moderate · 3 low
  • 47 Packagist advisories · 3 critical · 18 high · 21 moderate · 5 low
  • 38 Maven advisories · 12 critical · 11 high · 12 moderate · 3 low
  • 33 PyPI advisories · 3 critical · 9 high · 17 moderate · 4 low
  • 5 crates.io advisories · 2 critical · 1 high · 1 moderate · 1 low
  • 5 RubyGems advisories · 1 high · 3 moderate · 1 low
  • 4 NuGet advisories · 4 moderate
  • 1 Multi-ecosystem advisory · 1 high

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.