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 39 of 2026 · Retrospective · Covers 15 September 2026 11:00 to 22 September 2026 11:00 UTC

Read on the web

Top developments

Cisco warns of max severity ISE zero-day exploited in attacks

BleepingComputer · Incident · Remediation

Google fixes actively exploited Android zero-day on Pixel devices

BleepingComputer · Incident · Remediation

CISA orders feds to patch Zyxel flaw exploited for data theft

BleepingComputer · Advisory · Remediation

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

CISA · Advisory · Agentic SDLC

Malicious npm packages evade install-script defenses at runtime

BleepingComputer · Analysis · SCA

SAST

C-to-Rust Fallacy: Automatic Refactoring != Memory Security

arXiv · Research · SAST

SoK: From Finding to Deployment: Systematizing the OS Kernel Bug Lifecycle

arXiv · Research · DAST, Remediation, SAST

Auditing in the age of (good enough) AI

Trail of Bits · Analysis · Agentic SDLC, SAST

A Security Risk Assessment Framework for AI-Powered Development Tools

arXiv · Research · Agentic SDLC, SAST

SCA

Be alert: targeted attacks on prominent Rustaceans

Simon Willison · Incident · SCA

Brevo supply-chain attack injected ClickFix scripts on customer sites

BleepingComputer · Incident · SCA

KEVGraph: Exploitation-Aware Dependency Vulnerability Remediation

arXiv · Research · SCA

An undercover Google analyst infiltrated a notorious supply-chain hacking gang

Ars Technica: Security · Analysis · SCA

DAST

SyzHarness: Patch-Based Kernel Bug Reproduction with LLM-Synthesized Fuzzing Harnesses

arXiv · Research · DAST, Remediation

POZZER: A Power Side Channel-guided Fuzzer for Black-Box Embedded Systems

arXiv · Research · DAST

From Capability to Assurance in Autonomous Penetration-Testing Harnesses: A Framework and Reference Implementation

arXiv · Research · DAST

AIJon: Automated Generation of Annotations for Fuzzing

arXiv · Research · DAST

Agentic SDLC

Researchers used Anthropic’s Claude to hack into OpenAI

TechCrunch: Security · Research · Agentic SDLC

Researchers escape OpenAI Codex sandbox to run commands on host

BleepingComputer · Research · Agentic SDLC

BragJack attacks hijack AI browser agents through malicious extensions

BleepingComputer · Research · Agentic SDLC

Evaluating Coding Agents on Kernel Exploit Generation

arXiv · Research · Agentic SDLC

Remediation

CISA alerts of active exploitation of three Linux kernel flaws

BleepingComputer · Analysis · Remediation

CISA: Critical VMware RCE flaw now exploited by ransomware gangs

BleepingComputer · Incident · Remediation

Researchers use AI to find widespread software decoder flaw

CyberScoop · Research · Remediation

Microsoft fixes broken Excel copy and paste for all Office users

BleepingComputer · Analysis · Remediation

Trust

Zero Trust for AI Systems: Why Authorization Can’t Live Inside the Model

CoSAI · Research · Trust

Grow CRA Readiness: Find Your Path Through the European Union Cyber Resilience Act

OpenSSF · Standard / policy · Trust

ActGov: Governing LLM Agent Actions via Policy-Constrained Validation

arXiv · Research · Trust

Structured Decomposition for Reliable LLM-Generated Access Control Policies

arXiv · Research · Trust

Models

OPBackdoor: Opportunistic Backdoors via Alibi-Aligned Reasoning

arXiv · Research · Models

Decoding Guardrails: XAI-Guided Perturbation Analysis of Prompt Injection Detection

arXiv · Research · AI Security, Models

Context-Aware Pre-Deployment Evaluation of AI Systems: A Regulatory Framework for Nigerian Fintech

arXiv · Research · Models, Trust

Trustworthy Agentic AI: Failure Modes, Mitigation Strategies, and a Lifecycle Framework for Autonomous LLM Systems

arXiv · Research · AI Security, Models, Trust

AI Security

Spain's data agency gets first report of AI-powered data breach

BleepingComputer · Incident · AI Security

Building an AI Detection Engine That Understands Agent Intent

Wiz · Research · AI Security

Indirect tipping: a social attack surface in AI agent populations

arXiv · Research · AI Security

DUMA-Bench: A Dual-Control Multi-Agent Benchmark for Evaluating LLM Agent Security

arXiv · Research · AI Security

Advisories in this window

  • 11776 npm advisories · 11756 critical · 13 high · 7 moderate
  • 49 PyPI advisories · 24 critical · 15 high · 10 moderate
  • 42 Go advisories · 4 critical · 23 high · 13 moderate · 2 low
  • 35 Packagist advisories · 4 critical · 14 high · 15 moderate · 2 low
  • 26 Maven advisories · 2 critical · 13 high · 8 moderate · 3 low
  • 8 NuGet advisories · 1 critical · 5 high · 2 moderate
  • 6 crates.io advisories · 5 high · 1 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.