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

Daily, Saturday 19 September 2026 · Retrospective · Covers 18 September 2026 00:00 to 19 September 2026 00:00 UTC

Read on the web

Top developments

CVE-2026-53266: Linux Kernel Out-of-Bounds Write Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory · Agentic SDLC

Researchers use AI to find widespread software decoder flaw

CyberScoop · Research · Remediation

WeaselBiscuit Stealer Spreads via 13 npm Packages to Harvest Chrome Extension Storage

The Hacker News · Research · SCA

SAST

Auditing in the age of (good enough) AI

Trail of Bits · Analysis · Agentic SDLC, SAST

SCA

Micro-Collaborative Poisoning: A Distributed Attack on RAG Systems

arXiv · Research · SCA

TrustBOM: A Scalable Architecture for Confidentiality-Preserving SBOMs Across Organizations

arXiv · Research · SCA

Claimed Bug Bounty Hunter Likely Used LLM to Build PhantomRaven npm Stealer

The Hacker News · Analysis · SCA

Agentic SDLC

Defusing Explosive Prompts: Understanding and Preventing Trigger-Based Prompt Injections in LLM Agents

arXiv · Research · Agentic SDLC, AI Security

Plugin4Shell Lets Repository Owners Swap Pinned Plugin Code Across Four AI Coding Agents

The Hacker News · Analysis · Agentic SDLC, Remediation

Remediation

Microsoft fixes broken copy and paste for Excel 2016 users

BleepingComputer · Analysis · Remediation

New WordPress Click2Shell Flaw Forces Theme Installs, Can Chain to Code Execution

The Hacker News · Analysis · Remediation

Trust

A Governance-Aware Large Language Model Orchestrated Agentic Digital Twin for Transmission System Operator Control Room Decision Support

arXiv · Research · Trust

OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing

arXiv · Research · Trust

Models

End-to-End Hard-Label Cryptanalytic Model Extraction Using Efficient Sign Recovery

arXiv · Research · Models

CASCADE Against Jailbreaks: Combination Across Stages with Controlled Attack-Defense Evaluation

arXiv · Research · Models

HE-Guardrail: A Homomorphic Guardrail Against Jailbreak Attacks for Encrypted Large Language Model Inference

arXiv · Research · Models

AI Security

Zero-Trust Authorization and Discovery for Enterprise MCP

arXiv · Research · AI Security

A Vault with a Heap-View: The Uncomfortable Space Between AgentCore Harness and Identity

Palo Alto Networks · Analysis · AI Security

Advisories in this window

  • 29 npm advisories · 25 critical · 3 high · 1 moderate
  • 17 Go advisories · 1 critical · 9 high · 7 moderate
  • 12 PyPI advisories · 8 critical · 3 high · 1 moderate
  • 10 Packagist advisories · 2 high · 8 moderate
  • 2 Maven advisories · 1 critical · 1 high
  • 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.