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 41 of 2024 · Retrospective · Covers 1 October 2024 11:00 to 8 October 2024 11:00 UTC

Read on the web

Top developments

Qualcomm patches high-severity zero-day exploited in attacks

BleepingComputer · Advisory · Remediation

A Single Cloud Compromise Can Feed an Army of AI Sex Bots

Krebs on Security · Incident · Models

Recently patched CUPS flaw can be used to amplify DDoS attacks

BleepingComputer · Advisory · Remediation

Attackers exploit critical Zimbra vulnerability using cc’d email addresses

Ars Technica: Security · Research · Remediation

SAST

StagedVulBERT: Multi-Granular Vulnerability Detection with a Novel Pre-trained Code Model

arXiv · Research · SAST

SCA

Qualcomm Urges OEMs to Patch Critical DSP and WLAN Flaws Amid Active Exploits

The Hacker News · Advisory · Remediation, SCA

Securing the software supply chain with the SLSA framework

Trail of Bits · Standard / policy · SCA, Trust

Mitigating Downstream Model Risks via Model Provenance

arXiv · Research · SCA, Trust

PyPI Repository Found Hosting Fake Crypto Wallet Recovery Tools That Steal User Data

The Hacker News · Analysis · SCA

DAST

AI-Enhanced Ethical Hacking: A Linux-Focused Experiment

arXiv · Research · DAST

AutoPenBench: Benchmarking Generative Agents for Penetration Testing

arXiv · Research · DAST

LightSC: The Making of a Usable Security Classification Tool for DevSecOps

arXiv · Research · DAST

Remediation

Thousands of Linux systems infected by stealthy malware since 2021

Ars Technica: Security · Analysis · Remediation

Trust

Developing Assurance Cases for Adversarial Robustness and Regulatory Compliance in LLMs

arXiv · Research · Models, Trust

Strategic Insights from Simulation Gaming of AI Race Dynamics

arXiv · Research · Trust

Measurement challenges in AI catastrophic risk governance and safety frameworks

arXiv · Research · Trust

Models

SecAlign: Defending Against Prompt Injection with Preference Optimization

arXiv · Research · AI Security, Models

Defense-as-a-Service: Black-box Shielding against Backdoored Graph Models

arXiv · Research · Models

CAT: Concept-level backdoor ATtacks for Concept Bottleneck Models

arXiv · Research · Models

On the Adversarial Risk of Test Time Adaptation: An Investigation into Realistic Test-Time Data Poisoning

arXiv · Research · Models

AI Security

Agent Security Bench (ASB): Formalizing and Benchmarking Attacks and Defenses in LLM-based Agents

arXiv · Research · AI Security

Adversarial Decoding: Generating Readable Documents for Adversarial Objectives

arXiv · Research · AI Security, Models

Enhancing data privacy with layered authorization for Amazon Bedrock Agents

AWS Security Blog · Analysis · AI Security

Advisories in this window

  • 25 Packagist advisories · 3 high · 16 moderate · 6 low
  • 19 npm advisories · 5 critical · 4 high · 9 moderate · 1 low
  • 11 Go advisories · 5 high · 5 moderate · 1 low
  • 8 Maven advisories · 3 critical · 1 high · 4 moderate
  • 3 Multi-ecosystem advisories · 1 high · 2 moderate
  • 2 PyPI advisories · 2 moderate
  • 2 crates.io advisories · 1 high · 1 moderate
  • 1 RubyGems 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.