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 32 of 2025 · Retrospective · Covers 29 July 2025 11:00 to 5 August 2025 11:00 UTC

Read on the web

Top developments

SonicWall firewall devices hit in surge of Akira ransomware attacks

BleepingComputer · Incident · Remediation

Apple patches security flaw exploited in Chrome zero-day attacks

BleepingComputer · Advisory · Remediation

AI-powered Cursor IDE vulnerable to prompt-injection attacks

BleepingComputer · Research · Agentic SDLC, AI Security, Remediation

Wiz Research Uncovers Critical Vulnerability in AI Vibe Coding platform Base44 Allowing Unauthorized Access to Private Applications

Wiz · Research · Agentic SDLC, Remediation

Hackers exploit SAP NetWeaver bug to deploy Linux Auto-Color malware

BleepingComputer · Analysis · Remediation

SAST

Automated Code Repair for C/C++ Static Analysis Alerts

arXiv · Research · SAST

JSidentify-V2: Leveraging Dynamic Memory Fingerprinting for Mini-Game Plagiarism Detection

arXiv · Research · SAST

Think Broad, Act Narrow: CWE Identification with Multi-Agent Large Language Models

arXiv · Research · SAST

NATLM: Detecting Defects in NFT Smart Contracts Leveraging LLM

arXiv · Research · SAST

SCA

An Empirical Validation of Open Source Repository Stability Metrics

arXiv · Research · SCA

Turning ChatGPT Codex Into A ZombAI Agent

Embrace The Red · Analysis · Agentic SDLC, AI Security, SCA

🎉 Celebrating Five Years of OpenSSF: A Journey Through Open Source Security

OpenSSF · Analysis · SCA

AI-Generated Malicious npm Package Drains Solana Funds from 1,500+ Before Takedown

The Hacker News · Research · SCA

DAST

Towards Effective Offensive Security LLM Agents: Hyperparameter Tuning, LLM as a Judge, and a Lightweight CTF Benchmark

arXiv · Research · DAST

PentestJudge: Judging Agent Behavior Against Operational Requirements

arXiv · Research · DAST

LLM-Assisted Model-Based Fuzzing of Protocol Implementations

arXiv · Research · Agentic SDLC, DAST

Autonomous Penetration Testing: Solving Capture-the-Flag Challenges with LLMs

arXiv · Research · DAST

Agentic SDLC

Cursor IDE: Arbitrary Data Exfiltration Via Mermaid (CVE-2025-54132)

Embrace The Red · Research · Agentic SDLC

AgentSight: System-Level Observability for AI Agents Using eBPF

arXiv · Research · Agentic SDLC

Cursor’s AI coding agent morphed ‘into local shell’ with one-line prompt attack

CyberScoop · Analysis · Agentic SDLC, AI Security, Remediation

Remediation

Android gets patches for Qualcomm flaws exploited in attacks

BleepingComputer · Advisory · Remediation

Project Zero disclosure policy change puts vendors on early notice

CyberScoop · Standard / policy · Remediation

Activation-Guided Local Editing for Jailbreaking Attacks

arXiv · Research · Models, Remediation

Critical Dahua Camera Flaws Enable Remote Hijack via ONVIF and File Upload Exploits

The Hacker News · Advisory · Remediation

Trust

Trustworthiness of Legal Considerations for the Use of LLMs in Education

arXiv · Research · Trust

Evolution of AI Agent Registry Solutions: Centralized, Enterprise, and Distributed Approaches

arXiv · Research · AI Security, Trust

From Commits to Confidence: Towards Stability-Informed Risk Assessment in Open Source Software

arXiv · Research · Trust

DIRF: A Framework for Digital Identity Protection and Clone Governance in Agentic AI Systems

arXiv · Research · Trust

Models

BDFirewall: Towards Effective and Expeditiously Black-Box Backdoor Defense in MLaaS

arXiv · Research · Models

BadBlocks: Low-Cost and Stealthy Backdoor Attacks Tailored for Text-to-Image Diffusion Models

arXiv · Research · Models

Highlight & Summarize: RAG without the jailbreaks

arXiv · Research · Models

Large Reasoning Models Are Autonomous Jailbreak Agents

arXiv · Research · Models

AI Security

Using the NANDA Index Architecture in Practice: An Enterprise Perspective

arXiv · Research · AI Security

Defend LLMs Through Self-Consciousness

arXiv · Research · AI Security

Quoting @himbodhisattva

Simon Willison · Analysis · AI Security

Data Security in Large Language Models: Risks, Defense, and Directions

arXiv · Research · AI Security, Models

Advisories in this window

  • 46 npm advisories · 36 critical · 8 high · 1 moderate · 1 low
  • 18 PyPI advisories · 3 critical · 3 high · 9 moderate · 3 low
  • 17 Go advisories · 2 critical · 5 high · 8 moderate · 2 low
  • 11 Maven advisories · 11 moderate
  • 7 Packagist advisories · 1 critical · 1 high · 3 moderate · 2 low
  • 2 crates.io advisories · 1 high · 1 moderate
  • 2 NuGet advisories · 2 moderate
  • 1 RubyGems 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.