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 16 of 2026 · Retrospective · Covers 7 April 2026 11:00 to 14 April 2026 11:00 UTC

Read on the web

Top developments

CVE-2026-34621: Adobe Acrobat and Reader Prototype Pollution Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory · Remediation

Anthropic's Project Glasswing - restricting Claude Mythos to security researchers - sounds necessary to me

Simon Willison · Research · Remediation

CVE-2020-9715: Adobe Acrobat Use-After-Free Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory · Trust

Analysis of one billion CISA KEV remediation records exposes limits of human-scale security

BleepingComputer · Research · Remediation

OpenAI rotates macOS certs after Axios attack hit code-signing workflow

BleepingComputer · Analysis · SCA

SAST

Verify Before You Fix: Agentic Execution Grounding for Trustworthy Cross-Language Code Analysis

arXiv · Research · Agentic SDLC, SAST

The Code Whisperer: LLM and Graph-Based AI for Smell and Vulnerability Resolution

arXiv · Research · SAST

VulWeaver: Weaving Broken Semantics for Grounded Vulnerability Detection

arXiv · Research · SAST

VCAO: Verifier-Centered Agentic Orchestration for Strategic OS Vulnerability Discovery

arXiv · Research · DAST, SAST

SCA

Behind the Scenes: How StepSecurity Detected and Helped Remediate the Largest npm Supply Chain Attack

StepSecurity · Incident · SCA

@velora-dex/sdk Compromised on npm: Malicious Version Drops macOS Backdoor via launchctl Persistence

StepSecurity · Incident · SCA

Our response to the Axios developer tool compromise

OpenAI News · Incident · SCA

Security and Resilience in Autonomous Vehicles: A Proactive Design Approach

arXiv · Research · SCA

DAST

Why Your Automated Pentesting Tool Just Hit a Wall

BleepingComputer · Analysis · DAST

Hackers or Hallucinators? A Comprehensive Analysis of LLM-Based Automated Penetration Testing

arXiv · Research · AI Security, DAST

Agentic SDLC

LLM-Redactor: An Empirical Evaluation of Eight Techniques for Privacy-Preserving LLM Requests

arXiv · Research · Agentic SDLC

Resilient Write: A Six-Layer Durable Write Surface for LLM Coding Agents

arXiv · Research · Agentic SDLC

Securing Vibe Coding and AI Coding Agents: An End-to-End Approach with StepSecurity

StepSecurity · Analysis · Agentic SDLC

Security Concerns in Generative AI Coding Assistants: Insights from Online Discussions on GitHub Copilot

arXiv · Research · Agentic SDLC

Remediation

Intent redirection vulnerability in third-party SDK exposed millions of Android wallets to potential risk

Microsoft · Research · Remediation

A Queueing-Theoretic Framework for Dynamic Attack Surfaces: Data-Integrated Risk Analysis and Adaptive Defense

arXiv · Research · Remediation, SCA

Fine-grained Approaches for Confidence Calibration of LLMs in Automated Code Revision

arXiv · Research · Remediation

Does Pass Rate Tell the Whole Story? Evaluating Design Constraint Compliance in LLM-based Issue Resolution

arXiv · Research · Remediation

Trust

Beyond Static Sandboxing: Learned Capability Governance for Autonomous AI Agents

arXiv · Research · Trust

Post-Deployment Accountability in AI Governance: A Cross-Regulatory Empirical Analysis of AI Incidents

arXiv · Research · Trust

Expanding the MCP Maintainer Team

Model Context Protocol · Analysis · Trust

Blockchain and AI: Securing Intelligent Networks for the Future

arXiv · Research · Trust

Models

DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection

arXiv · Research · AI Security, Models

Safety Training Modulates Harmful Misalignment Under On-Policy RL, But Direction Depends on Environment Design

arXiv · Research · Models

Scaling Exposes the Trigger: Input-Level Backdoor Detection in Text-to-Image Diffusion Models via Cross-Attention Scaling

arXiv · Research · Models

Compiling Activation Steering into Weights via Null-Space Constraints for Stealthy Backdoors

arXiv · Research · Models, SCA

AI Security

‘GrafanaGhost’ bypasses Grafana’s AI defenses without leaving a trace

CyberScoop · Research · AI Security

WebAgentGuard: A Reasoning-Driven Guard Model for Detecting Prompt Injection Attacks in Web Agents

arXiv · Research · AI Security

TEMPLATEFUZZ: Fine-Grained Chat Template Fuzzing for Jailbreaking and Red Teaming LLMs

arXiv · Research · AI Security, Models

Modality-Native Routing in Agent-to-Agent Networks: A Multimodal A2A Protocol Extension

arXiv · Research · AI Security

Advisories in this window

  • 203 npm advisories · 107 critical · 24 high · 62 moderate · 10 low
  • 89 PyPI advisories · 16 critical · 27 high · 41 moderate · 5 low
  • 76 Go advisories · 9 critical · 29 high · 33 moderate · 5 low
  • 43 Maven advisories · 3 critical · 19 high · 19 moderate · 2 low
  • 23 Packagist advisories · 7 high · 10 moderate · 6 low
  • 19 crates.io advisories · 3 critical · 3 high · 9 moderate · 4 low
  • 10 NuGet advisories · 4 high · 6 moderate
  • 6 RubyGems advisories · 2 critical · 3 high · 1 low

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.