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 7 of 2026 · Retrospective · Covers 3 February 2026 12:00 to 10 February 2026 12:00 UTC

Read on the web

Top developments

CISA flags critical SolarWinds RCE flaw as exploited in attacks

BleepingComputer · Incident · Remediation

Detecting backdoored language models at scale

Microsoft · Incident · Models

CISA warns of five-year-old GitLab flaw exploited in attacks

BleepingComputer · Incident · Remediation

Compromised dYdX npm and PyPI Packages Deliver Wallet Stealers and RAT Malware

The Hacker News · Incident · SCA

Docker Fixes Critical Ask Gordon AI Flaw Allowing Code Execution via Image Metadata

The Hacker News · Research · Remediation

SAST

Secure Code Generation via Online Reinforcement Learning with Vulnerability Reward Model

arXiv · Research · SAST

Persistent Human Feedback, LLMs, and Static Analyzers for Secure Code Generation and Vulnerability Detection

arXiv · Research · SAST

Reading Between the Code Lines: On the Use of Self-Admitted Technical Debt for Security Analysis

arXiv · Research · SAST

SCA

Hackers exploit critical React Native Metro bug to breach dev systems

BleepingComputer · Incident · SCA

Join Us at Open Source SecurityCon Europe 2026 in Amsterdam

OpenSSF · Analysis · SCA

Software Testing at the Network Layer: Automated HTTP API Quality Assessment and Security Analysis of Production Web Applications

arXiv · Research · SCA

Malicious packages for dYdX cryptocurrency exchange empties user wallets

Ars Technica: Security · Analysis · SCA

DAST

CTFExplorer: Evaluating LLM Offensive Agents Through Multi-Target Web CTF Benchmarking

arXiv · Research · DAST, Remediation

Does Programming Language Matter? An Empirical Study of Fuzzing Bug Detection

arXiv · Research · DAST

Agentic SDLC

SWE-AGI: Benchmarking Specification-Driven Software Construction with MoonBit in the Era of Autonomous Agents

arXiv · Research · Agentic SDLC, Trust

Cisco Donates Project CodeGuard to Coalition for Secure AI

CoSAI · Analysis · Agentic SDLC, AI Security

Cisco’s Donation of Project CodeGuard to CoSAI: A New Chapter in Securing AI-Generated Code

CoSAI · Analysis · Agentic SDLC, Trust

Patch-to-PoC: A Systematic Study of Agentic LLM Systems for Linux Kernel N-Day Reproduction

arXiv · Research · Agentic SDLC, Remediation

AI-BOM

Permissive-Washing in the Open AI Supply Chain: A Large-Scale Audit of License Integrity

arXiv · Research · AI-BOM

Remediation

Analysis of active exploitation of SolarWinds Web Help Desk

Microsoft · Analysis · Remediation

BeyondTrust warns of critical RCE flaw in remote support software

BleepingComputer · Analysis · Remediation

CISA orders federal agencies to replace end-of-life edge devices

BleepingComputer · Analysis · Remediation

LLM-discovered 0 days

Anthropic · Analysis · Remediation

Trust

Timing and Memory Telemetry on GPUs for AI Governance

arXiv · Research · Trust

Aegis: Towards Governance, Integrity, and Security of AI Voice Agents

arXiv · Research · Trust

Pluralism in AI Governance: Toward Sociotechnical Alignment and Normative Coherence

arXiv · Research · Trust

Partnership on AI to Host Enterprise AI Governance Forum with JPMorganChase Ahead of India AI Impact Summit

Partnership on AI · Analysis · Trust

Models

Stop Testing Attacks, Start Diagnosing Defenses: The Four-Checkpoint Framework Reveals Where LLM Safety Breaks

arXiv · Research · Models

A Behavioral Fingerprint for Large Language Models: Provenance Tracking via Refusal Vectors

arXiv · Research · Models

Benchmarking Knowledge-Extraction Attack and Defense on Retrieval-Augmented Generation

arXiv · Research · Models

Red-teaming the Multimodal Reasoning: Jailbreaking Vision-Language Models via Cross-modal Entanglement Attacks

arXiv · Research · Models

AI Security

Autonomous Action Runtime Management(AARM):A System Specification for Securing AI-Driven Actions at Runtime

arXiv · Research · AI Security

MUZZLE: Adaptive Agentic Red-Teaming of Web Agents Against Indirect Prompt Injection Attacks

arXiv · Research · AI Security

CausalArmor: Efficient Indirect Prompt Injection Guardrails via Causal Attribution

arXiv · Research · AI Security

Agent-Fence: Mapping Security Vulnerabilities Across Deep Research Agents

arXiv · Research · AI Security

Advisories in this window

  • 143 npm advisories · 103 critical · 24 high · 13 moderate · 3 low
  • 50 Go advisories · 6 critical · 19 high · 22 moderate · 3 low
  • 35 Packagist advisories · 1 critical · 18 high · 11 moderate · 5 low
  • 20 PyPI advisories · 4 critical · 7 high · 9 moderate
  • 15 crates.io advisories · 5 critical · 3 high · 6 moderate · 1 low
  • 9 Maven advisories · 1 critical · 2 high · 5 moderate · 1 low
  • 6 RubyGems advisories · 4 high · 1 moderate · 1 low
  • 2 Multi-ecosystem advisories · 1 critical · 1 moderate
  • 2 GitHub Actions advisories · 1 high · 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.