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 49 of 2024 · Retrospective · Covers 26 November 2024 12:00 to 3 December 2024 12:00 UTC

Read on the web

Top developments

LLM Flowbreaking

Simon Willison · Research · Agentic SDLC, AI Security, Models

Race Condition Attacks against LLMs

Schneier on Security · Incident · AI Security, Models

CVE-2023-45727: North Grid Proself Improper Restriction of XML External Entity (XXE) Reference Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory

SAST

Dynamic Taint Tracking using Partial Instrumentation for Java Applications

arXiv · Research · SAST

SCA

Enhancing Security in Third-Party Library Reuse -- Comprehensive Detection of 1-day Vulnerability through Code Patch Analysis

arXiv · Research · Remediation, SCA

DAST

HackSynth: LLM Agent and Evaluation Framework for Autonomous Penetration Testing

arXiv · Research · DAST

The Seeds of the FUTURE Sprout from History: Fuzzing for Unveiling Vulnerabilities in Prospective Deep-Learning Libraries

arXiv · Research · DAST

DeepSeek AI: From Prompt Injection To Account Takeover

Embrace The Red · Analysis · AI Security, DAST

TransferFuzz: Fuzzing with Historical Trace for Verifying Propagated Vulnerability Code

arXiv · Research · DAST

Remediation

Critical Flaw in ProjectSend Under Active Exploitation Against Public-Facing Servers

The Hacker News · Analysis · Remediation

Small number of vulnerabilities patched in last Android security update of 2024

CyberScoop · Analysis · Remediation

New Windows Server 2012 zero-day gets free, unofficial patches

BleepingComputer · Analysis · Remediation

Microsoft re-releases Exchange updates after fixing mail delivery

BleepingComputer · Analysis · Remediation

Trust

Learning to Forget using Hypernetworks

arXiv · Research · Models, Trust

Responsible AI Governance: A Response to UN Interim Report on Governing AI for Humanity

arXiv · Research · Trust

What AI evaluations for preventing catastrophic risks can and cannot do

arXiv · Research · Trust

The effect of compliance requirements on vulnerability management strategies

Help Net Security · Standard / policy · Remediation, Trust

Models

Jailbreak Defense in a Narrow Domain: Limitations of Existing Methods and a New Transcript-Classifier Approach

arXiv · Research · Models

Trust & Safety of LLMs and LLMs in Trust & Safety

arXiv · Research · AI Security, Models

Robust and Transferable Backdoor Attacks Against Deep Image Compression With Selective Frequency Prior

arXiv · Research · Models

Improved Large Language Model Jailbreak Detection via Pretrained Embeddings

arXiv · Research · AI Security, Models

AI Security

0xfreysa/agent

Simon Willison · Analysis · AI Security

FLARE: Toward Universal Dataset Purification against Backdoor Attacks

arXiv · Research · AI Security, Models

Advisories in this window

  • 343 npm advisories · 337 critical · 2 high · 4 moderate
  • 12 Go advisories · 2 high · 10 moderate
  • 9 Packagist advisories · 3 high · 6 moderate
  • 8 Maven advisories · 1 critical · 3 high · 3 moderate · 1 low
  • 6 PyPI advisories · 1 high · 3 moderate · 2 low
  • 5 RubyGems advisories · 5 low
  • 1 crates.io advisory · 1 moderate
  • 1 NuGet 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.