惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 叶小钗
V
Visual Studio Blog
雷峰网
雷峰网
J
Java Code Geeks
博客园 - 三生石上(FineUI控件)
人人都是产品经理
人人都是产品经理
MyScale Blog
MyScale Blog
H
Hackread – Cybersecurity News, Data Breaches, AI and More
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
B
Blog RSS Feed
C
Check Point Blog
博客园 - Franky
酷 壳 – CoolShell
酷 壳 – CoolShell
博客园 - 【当耐特】
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
V
V2EX
D
Docker
IT之家
IT之家
博客园 - 聂微东
腾讯CDC
U
Unit 42
Microsoft Security Blog
Microsoft Security Blog
The Cloudflare Blog

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Mini Shai-Hulud: A persistent supply-chain worm
Dwayne McDan · 2026-05-26 · via DEV Community

On April 29th, Aikido researchers detected multiple compromised Node.js packages in SAP's namespace today. The malware adapts to CI environments, steals GitHub personal access tokens, and uses them to self-propagate—a pattern consistent with recent supply-chain attacks.

Update: May 12, 2026

Between May 11th and May 12th, a new phase of this attack was identified with numerous compromised NPM. While the big picture of the attack remains unchanged, the methodology shifted slightly.

While the original attack focused on a limited number of enterprise-ready packages, this new iteration broadens the scope by targeting more than 300 packages in a single wave. Those packages are spread across various namespaces, the most notable ones being @opensearch-project/opensearch and @mistralai/mistralai.

The OpenSearch NPM package total of 1.3M weekly downloads. Its compromise is currently the most recent in this attack wave and occurred at 00:39 am GMT. The malicious versions were removed by 10:45 am GMT. Affected versions are 3.5.3, 3.8.0, and 3.7.0.

The Mistral package is downloaded more than 2M times per week. The compromise occurred on May 11th, around 10:30 pm GMT. The malicious package versions 2.2.2, 2.2.3, and 2.2.4 were removed by May 11th, 11:00 pm GMT.

It is worth noting that both those packages were compromised as a result of the supply chain attack rather than directly.

The most interesting point about this attack is that most of the affected packages relied on trusted publishing to protect their NPM package publication. However, because the malware specifically targeted CI/CD environments, it could exploit the OIDC integration to bypass this protection.

This wave of the attack used the Session secure messaging software as the main exfiltration channel.

How it works

  1. Credential harvesting: The payload reads CI environment variables and adjusts its behavior accordingly.
  2. Token exfiltration: If a GitHub personal access token (ghp_) is found locally, the malware exfiltrates encrypted secrets to the corresponding GitHub account.
  3. Fallback mechanism: If no token is present, the malware scans commit messages for the string 'OhNoWhatsGoingOnWithGitHub:', then decodes a double-base64-encoded payload hidden inside, and uses it to exfiltrate the encrypted secrets to this account.

The RSA keys used to encrypt the exfiltrated secrets are the same as the ones used last week in the @bitwarden/cli attack.

Exfiltration infrastructure

GitGuardian identified 7 commits containing exposed ghp_ tokens—all remain valid and active at 16h46 EST. The attacker used the stolen tokens to create public repositories, each named with Dune-themed keywords. Inside each repository:

  • A README.md file containing the string "A Mini Shai-Hulud has Appeared"
  • A results/ directory with a JSON file holding the encrypted payload, protected by a 4096-bit RSA key

Campaign Footprint

Our telemetry shows the attack's full scope:

  • 23 GitHub accounts used for exfiltration
  • 971 public repositories created with stolen tokens
  • Top 6 accounts created 936 repos (96%) — these are the accounts tied to the 7 compromised tokens

GitGuardian Interactive Demo