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, Friday 31 July 2026 · Retrospective · Covers 30 July 2026 00:00 to 31 July 2026 00:00 UTC

Read on the web

Top developments

Microsoft Teams vishing attacks lead to Chaos ransomware attacks

BleepingComputer · Incident

CosmosEscape: Taking Over Every Database in Azure Cosmos DB

Wiz · Analysis · Remediation

Hackers Exploit AnySign4PC via Hacked Korean Sites to Install Backdoors Without Prompts

The Hacker News · Incident

SCA

Emerging Challenges in Threat Modeling for GenAI-Augmented Systems: A View from the Trenches

arXiv · Research · SCA

Security of World-Model-Based Embodied AI: A Lifecycle of Threats, Defenses, and Evaluation

arXiv · Research · AI Security, Models, SCA

Don't Trust the AI Ecosystem: Analyzing Privacy Leakage in Compromised Open-Source Components

arXiv · Research · Models, SCA

Agentic SDLC

Change2Task: From Repository Changes to Executable Coding Agent Tasks and Environments

arXiv · Research · Agentic SDLC

Escaping Linux Sandboxes via PipeWire (CVE-2026-5674)

Embrace The Red · Analysis · Agentic SDLC

Secure at Inception: Announcing the Snyk Studio Integration for Snowflake Cortex Code

Snyk · Analysis · Agentic SDLC, SCA

Remediation

Validation Evidence in LLM Repair Agents: How Much of What Passes Actually Tests the Bug?

arXiv · Research · Remediation

Google says AI helped Chrome fix 1,072 security bugs in two releases

BleepingComputer · Analysis · Remediation

Russian Hackers Exploit Microsoft OWA Flaw to Keep Mailbox Access After Credential Rotation

The Hacker News · Analysis · Remediation

Models

Temporal Poisoning: Clean-Label Backdoors via Event Redistribution in SNNs

arXiv · Research · Models

TriShield: Zero-Utility-Loss Defense Against Privacy Backdoors in Federated Language Model Fine-Tuning via Orthogonal Gradient Projection and Optimizer State Entanglement

arXiv · Research · Models

Benign on Label, Malicious by Design: Clean-Label Dormant-to-Activated Backdoor via Machine Unlearning with Removable Camouflage

arXiv · Research · Models

AI Security

CyberNeuro: A Privacy-Preserving Agentic Workbench for Cohort-Scale Neuroimage and Clinical Data Analysis

arXiv · Research · AI Security

Piggybacking on Perception: Stealthy Concurrent Audio Prompt Injections against Multimodal LLM Agents

arXiv · Research · AI Security

ChemMat-AgentSafetyBench: Evaluating Long-Horizon Attacks and Defenses in Chemistry and Materials Agents

arXiv · Research · AI Security

Advisories in this window

  • 212 npm advisories · 211 critical · 1 high
  • 10 PyPI advisories · 5 critical · 4 high · 1 moderate
  • 7 RubyGems advisories · 1 critical · 2 high · 3 moderate · 1 low
  • 3 Go advisories · 1 high · 2 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.