AI-Generated Malware Disguised as npm Package Drains Cryptocurrency Wallets

AI-Generated Malware Targets Cryptocurrency Wallets

In a concerning development at the intersection of artificial intelligence and cybersecurity, researchers have uncovered a malicious npm package that appears to have been generated using AI. The package, named "@kodane/patch-manager," was designed to drain cryptocurrency wallets and managed to attract over 1,500 downloads before being taken down 1

1

Source: Hacker News

Deceptive Packaging and Functionality

The malicious package masqueraded as a legitimate tool, claiming to offer "license validation and registry optimization" for Node.js applications. However, upon closer inspection, security researchers from Safety discovered its true nature as an "Enhanced Stealth Wallet Drainer" 2

2

The malware's functionality is particularly cunning. It targets cryptocurrency wallets on Windows, macOS, and Linux systems, draining funds to a predefined address on the Solana blockchain. Interestingly, it leaves enough currency in the wallet to cover transaction fees, potentially delaying detection 1

1

AI Fingerprints in the Code

What sets this malware apart is the strong indication that it was generated using AI, specifically Anthropic's Claude model. Paul McCarty, Safety's head of research, pointed out several telltale signs:

1. Liberal use of emojis in the source code
2. Well-written and descriptive comments throughout the code
3. Extensive JavaScript console logging messages
4. README.md files formatted in a style consistent with Claude-generated markdown
5. Frequent use of the word "Enhanced," a known Claude habit 1

   1

   2

   2

Source: The Register

McCarty noted, "For some reason code generating AI platforms love to put emojis in source code. No developer that I know does this, unless they are 14" 1

1

Implications for Software Supply Chain Security

The discovery of this AI-generated malware raises significant concerns about software supply chain security. The package's professional appearance and well-written documentation could easily deceive developers and bypass conventional security measures 2

2

Of particular concern is the use of postinstall scripts, which run automatically after a package is installed. This creates a dangerous blind spot, especially in CI/CD environments where dependencies are updated routinely without direct human review 2

2

Related Stories

Malicious npm package targets Claude AI users, leaks own GitHub token in sloppy attack

Popular OpenAI Codex tool with 29,000 weekly downloads exposed in npm supply chain attack

AI-Generated Code Hallucinations: A New Frontier in Software Supply Chain Attacks

Fake AI Tools Spread Noodlophile Malware, Targeting Crypto Wallets and Sensitive Data

Spread and Impact

The malicious package was uploaded on July 28, 2025, and flagged as malicious about two days later. In that short time, it managed to attract over 1,500 downloads. While all versions have now been removed, the actual impact remains unclear as the number of unique IP addresses that downloaded the package is unknown 1

1

2

Future Implications

This incident highlights the growing threat of AI-assisted malware creation. As AI tools become more sophisticated and accessible, cybercriminals can potentially create more convincing and dangerous malware that can evade traditional detection methods 2

2

The cybersecurity community now faces the challenge of not only monitoring for known malware but also developing strategies to detect and mitigate increasingly polished, AI-assisted threats that exploit trusted ecosystems like npm 2

2