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, Tuesday 27 October 2026 · Retrospective · Covers 26 October 2026 00:00 to 27 October 2026 00:00 UTC

Read on the web

Top developments

litellm 1.107.0 Compromised on PyPI and Docker: Credential Stealer and Hourly Remote Code Loader

StepSecurity · Incident · SCA

Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang

Cloudflare: Security · Research · Remediation

PHP2Shell security flaws now sucessfully exploited in the wild

BleepingComputer · Incident

SAST

ElectroNG is now open source!

Doyensec · Analysis · SAST

SCA

Software Supply Chain Security Solution Brief

Endor Labs · Analysis · SCA

DAST

Former pentester used ARTEX AI to steal 8.2 million data records

BleepingComputer · Analysis · DAST

Agentic SDLC

Your Wish Is My Commit: Over-Compliant Refactoring in AI Coding Agents

arXiv · Research · Agentic SDLC

Context Is a Pipeline, Not a Window: Information Attrition in LLM Agent Harnesses

arXiv · Research · Agentic SDLC

Reva: Real-Time Vulnerability Detection through Contextual Probing of Large Language Models

arXiv · Research · Agentic SDLC

Remediation

VU#128947: COVESA vsomeip excessive memory allocation in reliable (TCP) endpoint buffer handling

CERT/CC · Analysis · Remediation

Critical 'HTTP2Shell' Brotli vulnerability triggers emergency patches

BleepingComputer · Analysis · Remediation

How Absorption Rate Reveals Vulnerability Management Health

Cloud Security Alliance · Analysis · Remediation

Trust

Atomic Security Concerns: Towards a Foundational Language for Security Knowledge Representation

arXiv · Research · Trust

Governing Enterprise AI Traffic at Scale: Design and Operation of a Policy-Driven, Multi-Vendor LLM Gateway on Envoy

arXiv · Research · Trust

From Representations to Computational Trajectories: Intrinsic Fingerprinting for Large Language Model Provenance Verification

arXiv · Research · Models, Trust

Models

SpecGuard: Detecting Multi-Turn LLM Jailbreaks via Specification-Based Runtime Monitoring

arXiv · Research · Models

V$^3$AST-Bench: Verifiable, Valid, and Versatile Safety Evaluation for Coding Agents

arXiv · Research · Models

The Target Agent Is Its Own Best Critic but a Poor Proposer: Self-Improved Red Teaming

arXiv · Research · Models

AI Security

Log-To-Leak: Prompt Injection Attacks on Tool-Using LLM Agents via the Model Context Protocol

arXiv · Research · AI Security

Then there is prompt injection, a malicious email, ...: The Wild Wide West of OpenClaw Agentic System Security

arXiv · Research · AI Security

Scope-Gap Injection: Exploiting Evaluator Channel Blind Spots in the OpenClaw Agent Security Competition

arXiv · Research · AI Security

Advisories in this window

  • 58 npm advisories · 58 critical
  • 7 PyPI advisories · 7 critical

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.