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, Saturday 24 October 2026 · Retrospective · Covers 23 October 2026 00:00 to 24 October 2026 00:00 UTC

Read on the web

Top developments

An alignment assessment of the UK AISI cybertesting incident

Anthropic · Incident · AI Security

HPE patches eight max severity RCE and auth bypass AOS flaws

BleepingComputer · Advisory · Remediation

Critical x402 SDK Flaw Could Let Malicious APIs Drain AI Agents’ Token Balances

The Hacker News · Advisory · AI Security

SAST

Trusting the Tool: Asymmetric Deference of LLM Vulnerability Auditors to Static-Analysis Evidence

arXiv · Research · Remediation, SAST

Agentic SDLC

Cross-Platform Application Migration to HarmonyOS: An Agentic Pipeline with Artifact-Derived Verification

arXiv · Research · Agentic SDLC

DeepAVR: Deep Analysis via Multi-Agent Collaboration for Automated Vulnerability Repair

arXiv · Research · Agentic SDLC, Remediation

Remediation

New OpenSSF Tech Talk: How Upstream Contribution Minimizes Technical Debt & Streamlines CRA Reporting

OpenSSF · Standard / policy · Remediation, Trust

PHP shares mitigation for new RCE vulnerability with no patch

BleepingComputer · Analysis · Remediation

Beyond Copying: Lifelong Smart Contract Security Task Construction via Protocol Mutation

arXiv · Research · AI Security, Remediation

Models

Interpretable sparse fine-tuning with TopKLoRA enables faithful circuit discovery

arXiv · Research · Models

On Task, Off Track: Benchmarking LLM and Agent Robustness to Near-Target Prompt Injection

arXiv · Research · AI Security, Models

Data-Efficient Backdoor Attacks on Code Models via Trigger-Neutralized Counterfactual Scoring

arXiv · Research · Models

AI Security

A Deterministic Security Runtime for LLM Agents on MPI and Batch-Scheduled HPC Systems: Consensus-Governed, Hardware-Rooted Supervision of Every Rank

arXiv · Research · AI Security

Spoofing the Toolbox: Optimizing Tool Descriptions to Deceive LLM Agents

arXiv · Research · AI Security

You gave your AI agent access, but who is giving it orders?

BleepingComputer · Analysis · AI Security

Advisories in this window

  • 56 PyPI advisories · 26 critical · 11 high · 13 moderate · 6 low
  • 48 npm advisories · 21 critical · 11 high · 11 moderate · 5 low
  • 12 Go advisories · 1 critical · 4 high · 7 moderate
  • 3 crates.io advisories · 1 critical · 1 high · 1 moderate
  • 3 Multi-ecosystem advisories · 1 high · 2 low
  • 3 Packagist advisories · 1 critical · 2 moderate

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.