← 返回列表

📖 [The CloudSecList] Issue 362

单站 CloudSecList 发布 Sun, 01 Nov 2026 00:00:00 +0000 入库 2026-11-18T07:13:49

| |

| |

| |

| |

| CloudSecList is a newsletter for busy professionals who want to keep up to date with cloud security and its intersection with AI. Hand-curated by Marco Lancini since 2019. |

| |

| |

| Sponsor When Dormant Tokens Become Attack Paths A dormant token or over-permissioned identity is a puzzle piece that can unlock sensitive data. Frontier AI makes that puzzle cheap for attackers to solve. A public portal leaks API endpoints, a dormant token sits in a repo, a service account has no IP restriction, and a local admin can authorize an integration with org-wide read. Each one is low severity. Together, they add up to one catastrophic outcome. Obsidian Security explains why connections between findings change what deserves attention first. Read why frontier AI changes access-risk priorities |

| |

| This week's articles Security Logging in Cloud Environments - Kubernetes How to design a state of the art security logging platform for Kubernetes clusters, managed (EKS, GKE) or self-managed. Reproducible Context May be the Future of Vulnerability Response and Remediation AI is scaling discovery. The "human glue" and tribal knowledge behind response & remediation can't keep up. We need better foundational tools here. Securing our frontier model training environment OpenAI's post which describes how they secure the environments used to train and evaluate frontier models. Tailscale SSH: Findings and fixes Tailscale SSH vulnerabilities (the "Tailworm" report) stemmed from post-authorization permission enforcement, including SSH forwarding, and require an existing authorized tailnet foothold. Fixed in v1.104.4, with four security bulletins. Users should update, review SSH policies, and consider check mode. A Security Constitution at Work A "Security Constitution" (12 principles, 11 lessons, 13 trust anchors) guides an LLM harness to find insecure-by-design cloud flaws, meaning missing controls. It produced 109 filings across three clouds, and the article walks through five GCP and Azure cases. Broccoli2shell: Everything You Need To Know About The Critical Brotli Decoder Vulnerability Detect and mitigate CVE-2026-111497, a critical Brotli decoder vulnerability enables remote code execution. Cross-request KV cache exposure in the GLM-5.3 Workers AI model, and our remediation with SGLang A configuration mismatch in GLM 5.3 on Workers AI triggered an SGLang bug that exposed KV cache pages across requests. Here is how Cloudflare addressed the issue. Gheist: Attempted Scaled Abuse of Stolen GitHub Tokens Stumbles An automated campaign is planting secret-dumping workflows in every repository that stolen GitHub tokens can reach. It uses no C2 server, is actively evolving, and has been significantly disrupted by recent GitHub security improvements. Automating a YubiKey for Agentic Logins Jon Oberheide automated YubiKey touches for agentic logins, replacing flaky phone-call 2FA on one portal. A headless browser relays FIDO challenges to a host, where a Raspberry Pi Pico drives a reed relay grounding the key's touch sensor. |

| |

| |

| Sponsor Meet Deckard: the first open source scanner for shadow agents You can’t secure agents you can’t see. Deckard is the industry’s first open source scanner for shadow agents. Scan endpoints, cloud, managed agent platforms, and network logs to keep a live inventory of agents across your organization. Anyone on your security team can run a scan and send results straight to your SIEM. Spot shadow agents, investigate rogue behavior, and act fast. With Deckard, you can gain full visibility into agents running across your environments and keep risk in check. Download on Github |

| |

| Tools rea Reverse engineer anything with agents, from app behavior down to native binaries. zed-claude-devcontainer A sandboxed Claude Code dev container for the Zed editor. You can also check out the companion blog post. tk Git signing and SSH for agents, with expressive policies and keys that never leave Turnkey. ai-dfir-toolkit A vendor-neutral collection of Sigma, YARA, and Suricata rules for detecting compromise of LLM applications, MCP servers, ML supply chains, AI infrastructure, AI-powered insider threats, and RAG/vector database attacks. |

| |

| From the cloud providers #GCP   Inside Agent Compute Engine: Building the compute foundation for the agentic era Agent Compute Engine gives each AI agent a hardware-isolated microVM with container-like speed and no in-sandbox credentials. An external Agent Gateway swaps in credentials, idle agents are checkpointed, and memory is deduplicated via DAX and copy-on-write. |

| |

| Thanks for reading! |
| If you found this newsletter helpful, I'd really appreciate if you could forward it to your friends and colleagues! 👌 If you have questions, comments, or feedback, let me know on Twitter (@lancinimarco / @CloudSecList), or at feedback.cloudseclist.com! Thanks, Marco |

| |

| |

| |

| | | |

| |

| |

| Forward | Forward |

| |

| |

| Twitter | Tweet |

| |

| |

| Share | Share |

| |

| |

| How did you like this issue of CloudSecList? 1 2 3 4 5 |

| |

| Unsubscribe from CloudSecList |

| Archives | View in browser | Sponsorship |

| © 2019-present, CloudSecList by Marco Lancini. |

原文链接:https://cloudseclist.com/issues/issue-362/