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, Wednesday 29 July 2026 · Retrospective · Covers 28 July 2026 00:00 to 29 July 2026 00:00 UTC

Read on the web

Top developments

We now have a better understanding how OpenAI hacked into Hugging Face

Ars Technica: Security · Incident · Remediation

Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident

Simon Willison · Incident · AI Security, SCA

Disrupting supply chain attacks on npm and GitHub Actions

GitHub · Incident · SCA

SAST

How we use /goal to find bugs in Patch the Planet

Trail of Bits · Analysis · Agentic SDLC, Remediation, SAST, Trust

Hybrid Analysis for Secure MCP Tool Use in LLM Agents

arXiv · Research · AI Security, SAST

SecDrift: Measuring Sector-Conditioned Security Drift in AI-Generated Code

arXiv · Research · SAST

SCA

npm publish-time malware scanning and dual-use metadata

GitHub · Analysis · SCA

Dependabot alerts on malicious packages across more ecosystems

GitHub · Analysis · SCA

GitHub Actions holds potentially malicious workflows for approval

GitHub · Analysis · SCA

DAST

KQFuzz: Knowledge-Guided Fuzzing for Quantum Libraries via Large Language Models

arXiv · Research · DAST

Agentic SDLC

TraceCoder: Explainable and Auditable Code Generation with Position-Key Snippet Versioning

arXiv · Research · Agentic SDLC

SkillGate: Cost Efficient Runtime Malicious Skill File Detection in Coding Agents

arXiv · Research · Agentic SDLC, AI Security, SCA

Do Context Files Help Coding Agents? A Two-Agent Ablation Study on Real Repositories

arXiv · Research · Agentic SDLC

Remediation

Accelerating CISA BOD 26-04 Vulnerability and Triage Activities through Wiz

Wiz · Analysis · Remediation

Trust

Why Public Service AI Governance Frameworks Risk Failing in the Age of General-Purpose AI: Lessons from Policing

arXiv · Research · Trust

ARCHER: Agentic Rule and Compliance Harness for Executable Regulations

arXiv · Research · Trust

Architectural Backdoors in Vision-Language Model Supply Chains via Representation Steering

arXiv · Research · Models, Trust

Models

RAGuard: A Layered Defense Framework for Retrieval-Augmented Generation Systems Against Data Poisoning

arXiv · Research · Models

Learning Implicit Causal World Models from Multi-Agent Demonstrations

arXiv · Research · Models

GPT-Red: Automated Red Teaming via Self-Play at Scale

arXiv · Research · AI Security, Models

AI Security

MCP specification 2026-07-28

Model Context Protocol · Release · AI Security

The risk hiding behind exposed MCP servers

Wiz · Analysis · AI Security

Advisories in this window

  • 67 npm advisories · 63 critical · 2 high · 2 moderate
  • 20 PyPI advisories · 7 critical · 12 high · 1 low
  • 16 Go advisories · 1 critical · 6 high · 8 moderate · 1 low
  • 5 RubyGems advisories · 2 high · 1 moderate · 2 low
  • 3 crates.io advisories · 1 critical · 2 moderate
  • 3 Packagist advisories · 1 critical · 2 high
  • 2 NuGet advisories · 2 high
  • 2 Multi-ecosystem advisories · 2 high
  • 2 Maven advisories · 2 high

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.