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

Daily, Thursday 23 July 2026 · Retrospective · Covers 22 July 2026 00:00 to 23 July 2026 00:00 UTC

Read on the web

Top developments

CISA orders urgent action on actively exploited Langflow RCE flaw

BleepingComputer · Incident · AI Security, Remediation

Adobe Chrome extension flaw let sites access private WhatsApp chats

BleepingComputer · Analysis · Remediation

Microsoft Azure DevOps MCP Flaw Lets Hidden PR Comments Hijack AI Review Agents

The Hacker News · Incident · Agentic SDLC, AI Security

NTT DATA Group cuts incident analysis to 30 minutes with Codex

OpenAI News · Incident · Agentic SDLC

Trojanized Newtonsoft.Json Fork Hides Game-Rigging Code in a Working Library

The Hacker News · Research · SCA

SCA

Releases now reject new files after 14 days

Python Package Index · Analysis · SCA

Planned Updates to the PyPI User Interface

Python Package Index · Analysis · SCA

DAST

Quoting Thomas Ptacek

Simon Willison · Analysis · DAST

The Ethics of Autonomous AI Agents for Offensive Security

arXiv · Research · AI Security, DAST

Know Your Agent: Reconnaissance-Driven Pentesting of AI Agents

arXiv · Research · Agentic SDLC, AI Security, DAST

Agentic SDLC

IssueTrojanBench: Benchmarking AI Coding Agents Against Malicious Issue Requests

arXiv · Research · Agentic SDLC, AI Security, Models

Cross-Model LLM Code Review: Should you use Claude to review Codex or vice versa?

arXiv · Research · Agentic SDLC

Security Vulnerability Patterns in AI-Generated Code: A Cross-Model Comparative Study

arXiv · Research · Agentic SDLC

Remediation

Next chapter: Restructuring GitHub’s bug bounty program

GitHub · Analysis · Remediation

New InfraTrust report reveals infrastructure flaws admins should patch first

BleepingComputer · Analysis · Remediation

Trust

How enterprise GenAI can amplify ransomware risk — and how to contain it

BleepingComputer · Guide · Trust

The Fastest Path to AI Adoption Runs Through Security

The Hacker News · Analysis · Trust

Models

Chained Attacks on Drone-Based Federated Learning: From Network Disruption to Device Impersonation

arXiv · Research · Models

Geometric Configurations of Perturbed Jailbreak Prompts

arXiv · Research · Models

JANUS: Foreseeing Latent Risk for Long-Horizon Agent Safety

arXiv · Research · Models

AI Security

OpenAI’s accidental cyberattack against Hugging Face is science fiction that happened

Simon Willison · Analysis · AI Security

Advisories in this window

  • 55 npm advisories · 15 critical · 21 high · 19 moderate
  • 22 Maven advisories · 12 high · 10 moderate
  • 9 PyPI advisories · 3 high · 3 moderate · 3 low
  • 6 Packagist advisories · 4 moderate · 2 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.