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 37 of 2026 · Retrospective · Covers 1 September 2026 11:00 to 8 September 2026 11:00 UTC

Read on the web

Top developments

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

BleepingComputer · Incident · Remediation

N-able patches max severity N-central flaw amid ongoing attacks

BleepingComputer · Advisory · Remediation

Unpatched Magento and Adobe Commerce Zero-Day Exploited to Backdoor Online Stores

The Hacker News · Incident · Remediation

OpenAI admits it didn't disclose rogue AI wiki hijacking incident

BleepingComputer · Incident · AI Security

39 New Methods That Compromise Passkey Authentication

BleepingComputer · Incident · AI Security, Trust

SAST

Testing race conditions with memory access tracing and stack-based delay injection

Google · Analysis · SAST

An Empirical Analysis of CodeQL False Positives and Query Refinements for Java Vulnerabilities

arXiv · Research · Remediation, SAST, Trust

Automated Vulnerability Injection in Smart Contracts Using Large Language Models

arXiv · Research · SAST

Towards Behavior Tree-Guided Vulnerability Detection with Lightweight LLMs

arXiv · Research · SAST

SCA

BGP hijack infecting networks caused by a comedy of errors that’s not funny at all

Ars Technica: Security · Incident · SCA

VEX-Bench: Benchmarking LLM Agents for Assessing Exploitability of Software Supply Chain Vulnerabilities

arXiv · Research · SCA, Trust

Do AI Coding Assistants Check Before They Install? A Pre-Registered Demand-Side Audit of Trust Signals in the Research Software Supply Chain

arXiv · Research · Agentic SDLC, SCA, Trust

Scanning the Harness: Configuration Exposures in AI Coding-Agent Supply Chains

arXiv · Research · Agentic SDLC, SCA

DAST

LLM-Based Penetration Testing in the Presence of Honeypots

arXiv · Research · DAST

Staying on the Attack Path: Structured State for Long-Horizon Automated Penetration Testing

arXiv · Research · DAST

Robustness-Aware Evaluation and Enhancement of Mutation-Based Fuzzing for Bug Discovery

arXiv · Research · DAST

Burp Suite vs OWASP ZAP: Which Fits Your Testing Workflow?

Aikido Security · Analysis · DAST

Agentic SDLC

Authority Is Not a String: A Capability-Scoped Harness for Prompt-Injection-Resistant Coding Agents

arXiv · Research · Agentic SDLC, AI Security

SWE-Bench Pro Verified: A Reliable Benchmark for Software Engineering Agents

arXiv · Research · Agentic SDLC, Models

Beyond Lexical Metrics: Sentence-Embedding Detection of Reviewer Habituation in AI Code Review

arXiv · Research · Agentic SDLC

Building a research-software catalog with a coding agent: from hackathon prototype to public deployment

arXiv · Research · Agentic SDLC

Remediation

Attackers exploit zero-days in consistently besieged SonicWall product

CyberScoop · Analysis · Remediation

Critical Elementor Pro flaw exploited to take over WordPress sites

BleepingComputer · Incident · Remediation

HPE patches critical ArubaOS-CX remote code execution flaw

BleepingComputer · Advisory · Remediation

ConnectWise warns of new ScreenConnect flaw without patch

BleepingComputer · Analysis · Remediation

Trust

US Becomes Top Target in RMM Phishing Campaign Spanning 46 Countries

The Hacker News · Incident · Trust

Seeing is Not Believing: Breaking the Physical-to-Digital Trust Boundary in Robotics

arXiv · Research · Trust

When Is Content "AI-Generated Enough"? Labelling Synthetic Media under the Digital Services Act and the AI Act

arXiv · Research · Trust

MOLE: Detecting Insider Threats in AI Agents

arXiv · Research · AI Security, Trust

Models

Trust-But-Verify: Poisoning-Resilient Locally Private Graph Learning Protocols

arXiv · Research · Models

TrojanWorld: Backdooring World-Model Agents via Imagination Steering

arXiv · Research · Models

An Empirical Measurement of Jailbreaking Evaluators

arXiv · Research · Models

The Geometry of Refusal: Why Post-Hoc Refusal Is Fragile and Pretraining-Time Refusal Persists

arXiv · Research · Models

AI Security

ASCII smuggling crosses over from AI prompt injection to phishing evasion

Microsoft · Analysis · AI Security

CoSAI Hosts Working Session on Open Delegation & Identity Standard (ODIS)

CoSAI · Standard / policy · AI Security

The hidden risks of shadow AI

UK National Cyber Security Centre · Analysis · AI Security

AgentDrift: A Step-Labeled Benchmark of Injection-Hijacked LLM Agent Trajectories

arXiv · Research · AI Security

Advisories in this window

  • 27959 npm advisories · 27916 critical · 25 high · 16 moderate · 2 low
  • 60 PyPI advisories · 34 critical · 10 high · 16 moderate
  • 49 Go advisories · 7 critical · 26 high · 16 moderate
  • 22 Packagist advisories · 10 high · 11 moderate · 1 low
  • 11 Multi-ecosystem advisories · 1 critical · 8 high · 2 moderate
  • 4 crates.io advisories · 1 high · 3 moderate
  • 2 RubyGems advisories · 1 high · 1 moderate
  • 2 Maven advisories · 2 high
  • 1 Hex advisory · 1 high
  • 1 NuGet advisory · 1 low

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.