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 20 of 2026 · Retrospective · Covers 5 May 2026 11:00 to 12 May 2026 11:00 UTC

Read on the web

Top developments

Palo Alto Networks warns of firewall RCE zero-day exploited in attacks

BleepingComputer · Incident · Remediation

DAEMON Tools trojanized in supply-chain attack to deploy backdoor

BleepingComputer · Incident · SCA

Mini Shai-Hulud Strikes Again: TanStack + more npm Packages Compromised

Wiz · Incident · SCA

Official CheckMarx Jenkins package compromised with infostealer

BleepingComputer · Incident · SCA

Flaw in Claude’s Chrome extension allowed ‘any’ other plugin to hijack victims’ AI

CyberScoop · Incident · Agentic SDLC

SAST

Safety Context Injection: Inference-Time Safety Alignment via Static Filtering and Agentic Analysis

arXiv · Research · Models, SAST

Benchmarking LLM-Based Static Analysis for Secure Smart Contract Development: Reliability, Limitations, and Potential Hybrid Solutions

arXiv · Research · SAST

MARGIN: Margin-Aware Regularized Geometry for Imbalanced Vulnerability Detection

arXiv · Research · SAST

DCVD: Dual-Channel Cross-Modal Fusion for Joint Vulnerability Detection and Localization

arXiv · Research · SAST

SCA

DAEMON Tools devs confirm breach, release malware-free version

BleepingComputer · Incident · SCA

Quasar Linux RAT Steals Developer Credentials for Software Supply Chain Compromise

The Hacker News · Incident · SCA

FlowSteer: Prompt-Only Workflow Steering Exposes Planning-Time Vulnerabilities in Multi-Agent LLM Systems

arXiv · Research · SCA

Under the Hood of SKILL.md: Semantic Supply-chain Attacks on AI Agent Skill Registry

arXiv · Research · AI Security, SCA

DAST

Detecting Web Server Probing & Fuzzing in Traefik with Automated Cloudflare Response

Elastic · Research · DAST

Cochise: A Reference Harness for Autonomous Penetration Testing

arXiv · Research · DAST

Continuous Discovery of Vulnerabilities in LLM Serving Systems with Fuzzing

arXiv · Research · DAST

From Controlled to the Wild: Evaluation of Pentesting Agents for the Real-World

arXiv · Research · DAST

Agentic SDLC

CVE-2026-0300: Palo Alto Networks PAN-OS Out-of-bounds Write Vulnerability added to CISA's Known Exploited Vulnerabilities catalog

CISA · Advisory · Agentic SDLC

Elastic Security MCP App: Interactive security operations inside your AI Tools

Elastic · Analysis · Agentic SDLC, AI Security

CrackMeBench: Binary Reverse Engineering for Agents

arXiv · Research · Agentic SDLC

Malicious Coding Agent Skills and the Risk of Dynamic Context

Datadog · Analysis · Agentic SDLC, AI Security

Remediation

A DOD contractor’s API flaw exposed military course data and service member records

CyberScoop · Research · Remediation

Demystifying and Detecting Agentic Workflow Injection Vulnerabilities in GitHub Actions

arXiv · Research · Remediation, SCA

Patch2Vuln: Agentic Reconstruction of Vulnerabilities from Linux Distribution Binary Patches

arXiv · Research · Remediation

Agentic AI and the Industrialization of Cyber Offense: Forecast, Consequences, and Defensive Priorities for Enterprises and the Mittelstand

arXiv · Research · AI Security, Remediation

Trust

Digital Identity for Agentic Systems: Toward a Portable Authorization Standard for Autonomous Agents

arXiv · Research · Trust

The Granularity Mismatch in Agent Security: Argument-Level Provenance Solves Enforcement and Isolates the LLM Reasoning Bottleneck

arXiv · Research · AI Security, Trust

Governing AI-Assisted Security Operations: A Design Science Framework for Operational Decision Support

arXiv · Research · Agentic SDLC, Trust

The Challenges of Balancing AI Compliance and Technological Innovations in Critical Sectors: A Systematic Literature Review

arXiv · Research · Trust

Models

On-Policy Self-Evolution via Failure Trajectories for Agentic Safety Alignment

arXiv · Research · Models

Persona-Conditioned Adversarial Prompting (PCAP): Multi-Identity Red-Teaming for Enhanced Adversarial Prompt Discovery

arXiv · Research · Models

LITMUS: Benchmarking Behavioral Jailbreaks of LLM Agents in Real OS Environments

arXiv · Research · Models

Re-Triggering Safeguards within LLMs for Jailbreak Detection

arXiv · Research · Models

AI Security

When prompts become shells: RCE vulnerabilities in AI agent frameworks

Microsoft · Research · AI Security

IPI-proxy: An Intercepting Proxy for Red-Teaming Web-Browsing AI Agents Against Indirect Prompt Injection

arXiv · Research · AI Security

Mobius Injection: From One-Shot Prompt Injection to Sustained Resource Exhaustion in AI Agent Infrastructure

arXiv · Research · AI Security

Options, Not Clicks: Lattice Refinement for Consent-Driven MCP Authorization

arXiv · Research · AI Security

Advisories in this window

  • 285 npm advisories · 204 critical · 44 high · 34 moderate · 3 low
  • 145 PyPI advisories · 13 critical · 62 high · 57 moderate · 13 low
  • 117 Go advisories · 17 critical · 39 high · 51 moderate · 10 low
  • 97 Packagist advisories · 11 critical · 36 high · 47 moderate · 3 low
  • 47 Maven advisories · 6 critical · 19 high · 17 moderate · 5 low
  • 41 crates.io advisories · 4 critical · 17 high · 15 moderate · 5 low
  • 12 Hex advisories · 6 high · 4 moderate · 2 low
  • 8 NuGet advisories · 5 high · 3 moderate
  • 7 RubyGems advisories · 1 high · 6 moderate
  • 3 Multi-ecosystem advisories · 2 high · 1 moderate
  • 1 Swift advisory · 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.