Daily Edition · Page 5

THE DAILY INDEX

Technology, AI, software, security, crypto and business

ATTACKS & VULNERABILITIES

TLDR InfoSec

VANHELSING RANSOMWARE BUILDER LEAKED ON HACKING FORUM

After a former developer for the Ransomware-as-a-Service operator VanHelsing attempted to sell the source code for the VanHelsing affiliate panel, data leak Tor sites, and builders for the Windows and Linux encryptors, VanHelsing proactively released the source code itself. The leaked source code includes the Windows encryptor, which will require modifications for other threat actors, and the source code for the affiliate panel and blog site, which can be run as is. The data VanHelsing leaked was only a subset of what the former developer attempted to sell, as it did not include any databases or the Linux builder.

WHY RECREATING AN IAM ROLE DOESN'T RESTORE TRUST: A GOTCHA IN ROLE ARNS

In AWS IAM, trust policies reference ARNs in their JSON but reference an underlying, immutable principal ID, so if a role is deleted and recreated, the trust policy will also need to be recreated, even if the ARN is unchanged. This provides a security benefit in that if a role is deleted but the trust policy is unintentionally not updated, the role cannot be stealthily recreated and assume the same level of trust. Users can get around this issue by using a complete account as the principal in a trust policy with a condition block that restricts access to the specific role ARN.

FORGET IPS: USING CRYPTOGRAPHY TO VERIFY BOT AND AGENT TRAFFIC

In the era of AI agents, not all bot traffic is malicious. Cloudflare proposed two mechanisms for “good bots” to self-identify. The first is HTTP Message Signatures, which adds a Signature-Input header with a signature, validity period, usage tag, and a Signature-Agent header to locate the key. The second option is a new TLS flag, req_mTLS, which allows clients to indicate support for mutual TLS and identify the bot.

TLDR IndexPage 5