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 38 of 2026 · Retrospective · Covers 8 September 2026 11:00 to 15 September 2026 11:00 UTC

Read on the web

Top developments

CVE-2026-20079: Cisco Firewall Management Center Authentication Bypass Using an Alternate Path or Channel Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory · Remediation

Cisco patches Secure Email Gateway zero-day exploited in attacks

BleepingComputer · Advisory · Remediation

Google warns of new Chrome zero-day bug exploited in attacks

BleepingComputer · Incident · Remediation

New Microsoft Defender 'ShieldCrash' zero-day grants SYSTEM access

BleepingComputer · Research · Remediation

CVE-2026-87491: Google Chromium V8 Out of Bounds Write Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory · Agentic SDLC

SAST

Vulnerability Localization Benchmark: Measuring Agentic Security Analysis at Repository Scale

arXiv · Research · AI Security, Remediation, SAST

Automating Attack Graph Construction for Agentic Pentesting. Towards Neuro-Symbolic Vulnerability Hunting

arXiv · Research · Agentic SDLC, DAST, SAST

EchoFuzz: Empowering Smart Contract Fuzzing with Large Language Models

arXiv · Research · DAST, SAST

What is the Difference Between Me and You? Benchmarking the Quality Gap Between Human-Written and AI-Generated Code

arXiv · Research · Agentic SDLC, SAST

SCA

Twitch Browser Extension Leaks OAuth Tokens From Nearly 31,000 Users

The Hacker News · Incident · SCA

An Exploratory Study of Dependabot Cooldown Adoption in Open-Source GitHub Projects

arXiv · Research · SCA

Homebrew 7.0.0 gets built-in GUI, better security controls

BleepingComputer · Analysis · SCA

Empowering Open Source Security with Scalable Infrastructure

OpenSSF · Analysis · SCA

DAST

Toward Secure AI-Powered Penetration Testing Agents: Security Threats, Guardrails, and Architectural Perspectives

arXiv · Research · AI Security, DAST

IntentFuzz: A Protocol-Aware Fuzzer for Automated Invariant Violation Detection in Intent-Based Cross-Chain Bridges

arXiv · Research · DAST

Big Enough to Break Out: Tracking the Rising Capability of LLM Penetration-Testing Agents

arXiv · Research · DAST

Why federal cyber defense demands an offense-driven mindset

CyberScoop · Analysis · DAST

Agentic SDLC

Assurance Envelopes for Autonomous Coding Agents: Minimum-Cost Evidence for Software Change

arXiv · Research · Agentic SDLC

AgentGuard: Learning Execution Guardrails from Anomalous Coding-Agent Trajectories

arXiv · Research · Agentic SDLC

Adversarial Testing of Automated Program Repair Agents for Security Vulnerabilities

arXiv · Research · Agentic SDLC

Using Agentic AI for contextualized and multifaceted code review at Ericsson

arXiv · Research · Agentic SDLC

AI-BOM

SPDX Articles in Journal of Innovation

SPDX · Analysis · AI-BOM, SCA, Trust

Remediation

Microsoft September 2026 Patch Tuesday fixes 966 flaws, 2 zero-days

BleepingComputer · Advisory · Remediation

CISA Flags Exploited Cisco, Citrix, Fortinet Flaws, Sets Sept. 12 Federal Patch Deadline

The Hacker News · Incident · Remediation

PaperCut Replaces Emergency Patches With Fixes for Two Actively Exploited Flaws

The Hacker News · Advisory · Remediation

SAP warns of maximum severity 'OVERPASS' kernel vulnerability

BleepingComputer · Analysis · Remediation

Trust

The EU CRA's Real Question: What Shipped, and When Did You Know?

BleepingComputer · Standard / policy · Trust

Mapping U.S. Federal AI Governance Against Sector Vulnerability

arXiv · Research · Trust

Authorization Architectures for Tool-Using AI Agents

arXiv · Research · AI Security, Trust

Toward Sustainable AI Deployment: A Carbon-Aware Decision Framework for Enterprise Supply Chain Systems

arXiv · Research · Trust

Models

Verbalizing Subliminal Learning Effects Using Text Optimization

arXiv · Research · Models

MarkSec: Capability-Aware Evaluation of Adversarial Attacks Against LLM Watermarks

arXiv · Research · Models

Implementing a White-Box Undetectable Backdoor for Random Fourier Features

arXiv · Research · Models

When the World Lies: Backdoor Attacks on Latent World Models for Downstream Control

arXiv · Research · Models

AI Security

AI-powered attack exploited PaperCut flaws to hack 395 organizations

BleepingComputer · Incident · AI Security

An alignment assessment of recent cybersecurity incidents

Anthropic · Research · AI Security

Anthropic Discloses Fourth AI Hacking Incident Involving Claude Opus 4.6

The Hacker News · Incident · AI Security

Universal Defenses for Tool-Integrated LLM Agents Against Adversarial Attacks

arXiv · Research · AI Security, Models

Advisories in this window

  • 27582 npm advisories · 27511 critical · 40 high · 28 moderate · 3 low
  • 73 PyPI advisories · 25 critical · 23 high · 25 moderate
  • 38 Go advisories · 4 critical · 12 high · 19 moderate · 3 low
  • 17 Packagist advisories · 2 critical · 9 high · 3 moderate · 3 low
  • 7 Maven advisories · 3 critical · 2 high · 2 moderate
  • 7 NuGet advisories · 1 critical · 4 high · 2 moderate
  • 4 crates.io advisories · 1 critical · 2 high · 1 moderate
  • 1 RubyGems advisory · 1 moderate
  • 1 Multi-ecosystem 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.