TLDR InfoSec

TLDR INFORMATION SECURITY 2026-10-21

6 sections20 entrieslinks blockcomplete parse

Parse notes

  • Subscription/account footer truncated before parsefooter_truncated

01GENERAL

Sponsored

BLOCK A SIMULATED ATTACK WITH FORTIGATE ON AWS

Traffic moving between your AWS Region and Outposts can bypass inspection if your security stack isn't positioned in the middle. This solution template deploys a fully configured FortiGate Next-Generation Firewall [5] with AWS Gateway Load Balancer in one AWS CloudFormation stack. What you get: • Automated deployment: VPCs, Transit Gateway, IAM roles, and security policies — pre-configured. • Unified hybrid inspection: Monitor north-south and east-west traffic across VPCs and Outposts from one dashboard. • Hands-on IPS test: Create a sensor and block a simulated malicious download, step by step.

aws.amazon.comSponsor

02ATTACKS & VULNERABILITIES

HACKERS USE NEW ROMULUS CLICKFIX PLATFORM TO DELIVER LEGIONLOADER AND OTHER MALWARE

ROMULUS, a ClickFix delivery platform, was discovered using fake landing pages to make users run malicious commands. It was found to serve PowerShell staging scripts that deploy trojanized MSI installers or padded encrypted archives containing Legion Loader. Researchers found six IP addresses running the platform. One also hosted a ClickFix script linked to malicious ChatGPT Custom GPTs and Legion Loader, then retrieved further payloads.

thehackernews.com2 min read

META MUSE EXPOSES USERS' PERSONAL AND BUSINESS SECRETS WHEN THEY SHARE CONVERSATIONS

Researchers at Claroc found 6,800 Meta Muse share links through passive DNS searches. The links included conversation titles in their subdomains, making them searchable. Samples exposed executive-search notes, revenue reviews, loan applications, bank-fraud details, property records, and staff calendars. Meta paused the sharing feature on October 20 while reviewing the report.

arstechnica.com3 min read

BEYONDTRUST REMOTE SUPPORT 26.2.3 SUFFERS AUTH BYPASS CVE-2026-102764

BeyondTrust Remote Support versions 26.2.3 and earlier are affected by CVE-2026-102764, a critical authentication bypass and arbitrary file upload vulnerability with a CVSS score of 9.3. CVE-2026-102764 lets unauthenticated attackers perform file operations, escalate to root, and access credential vaults. A suspected Chinese state-sponsored espionage group, Armory Castle (UNC5221), has actively exploited this zero-day flaw since late August to deploy custom webshells such as BARECLAW for persistence and lateral movement. BeyondTrust released security patches in October for self-hosted customers, prompting CISA and the Canadian Centre for Cyber Security to advise organizations to restrict external access to appliances and conduct forensic reviews for signs of compromise.

links.tldrnewsletter.com4 min read

03STRATEGIES & TACTICS

INSIDE CAIRN: OPEN WINDOWS INTO MASSIVE AI-DRIVEN CYBER OPERATIONS

NorForge found more than 20 publicly exposed Cairn panels used for malicious intrusion work. One Chinese-language cluster ran over 1,100 attack projects against e-commerce and payment systems in the US, Japan, Australia, and Canada, seeking unauthenticated RCE, file-upload webshells, SQL injection, and authentication bypasses. Another cluster targeted merchant payment, crypto, and pharmaceutical sites for card and cryptocurrency theft. The panels exposed targets, credentials, API keys, agent instructions, and successful compromises. Trufflehog found hundreds of secrets, with 11% verified valid. NorForge is notifying victims and seeking panel takedowns.

norforge.com16 min read

PACKAGE MANAGER SANDBOXING

Homebrew developer Andrew Nesbitt explains Homebrew 7.0's split fetch and install phases as well as the Landlock and Seatbelt-based sandboxes. Nesbitt then reviews the sandbox implementations, allowlist implementations, and sandbox proposals from other package managers. Some package managers like Go and Elm go a step further by fully removing code execution.

nesbitt.io7 min read

LAMBDA WEB FUNCTIONS: PUBLIC BY DEFAULT, AND ALREADY IN YOUR LAMBDA:* GRANTS

Lambda Web Functions are a new AWS feature that allow for running a Node.js HTTP server behind a managed HTTPS domain. By default, the functions run with no authentication, and role credentials are included as environment variables, which could potentially be leaked for the full duration of the execution environment. Existing roles that grant lambda:* permissions have access to web functions by default, existing SCPs may not cover them, and visibility tooling may also miss functions, leading to the potential for shadow Web Functions being created and going unnoticed.

act.security5 min read

04LAUNCHES & TOOLS

Sponsored

ADDRESS API AND AGENTIC AI RISKS

Discover how excessive data exposure, tool poisoning, privilege escalation, and shifting trust boundaries affect modern architectures. Hear practical guidance for embedding security early, improving token hygiene, and preparing for regulatory scrutiny. Watch the webinar on

pages.awscloud.comSponsor

MIND (PRODUCT LAUNCH)

MIND is a data loss prevention platform that detects and blocks data exfiltration across email, endpoints, generative AI tools, and SaaS. It classifies content and context, then investigates events, tunes policies, and resolves issues.

mind.io

QUICKSAND

Quicksand is an async Python API for launching, controlling, and snapshotting QEMU virtual machines with a particular focus on sandboxing AI agents.

github.comGitHub repository

VELOCIRAPTOR SKILLS

A collection of public, reusable Codex skills and DFIR tooling for Velociraptor setup, collection, hunting, and host analysis.

github.comGitHub repository

05MISCELLANEOUS

A HACKER BROKE INTO JOHN DEERE AND SHARKNINJA WITH AN ARMY OF AI AGENTS

A hacker used autonomous agents on the open-source Cairn platform to breach more than 100 sites, reportedly including John Deere, SharkNinja, and Cummins. The attacker supplied intrusion tactics in under 100 Chinese characters. The breaches took about 23 hours. Norforge found over 20 likely malicious Cairn instances. John Deere said no sensitive data was accessed. SharkNinja contained the incident.

theverge.com2 min read

DHS TO BAN DEEPSEEK FROM FEDERAL CIVILIAN AGENCIES

The Department of Homeland Security announced a new ban on the procurement of products from DeepSeek. The directive cited national security concerns stemming from the potential for sensitive data to be stored or processed in China. Agencies will have 60 days to request exceptions to the policy.

cyberscoop.com2 min read

OUR LATEST DEFENSES AGAINST EVOLVING WEB SCAMS

Chrome's Enhanced Protection added ClickFix detection that scrubs PII from copied commands, sends them to Safe Browsing for analysis, and warns users when the verdict comes back dangerous. Google also moved its Gemini scam detection from on-device to a server-side model that scores page intent and brand impersonation, and it added a lower-confidence "suspicious site" warning tier beneath the familiar red interstitial. The company claimed Enhanced Protection users were ten times safer than Standard Protection users, a figure drawn from Google's own measurement.

blog.google3 min read