Just saw a rather unsettling security report, and I want to discuss this topic with everyone.

Recently, security researchers discovered over 300 malicious AI plugins stealing user data, especially targeting crypto users. This is not a small issue — it means that those seemingly harmless AI assistants on your computer could actually have become backdoors for hackers to access your accounts.

Many people are now using local AI assistants to help organize files, analyze transactions, handle emails, and even directly connect wallets and trading tools. Sounds convenient, right? The problem is, once these AI tools are embedded with malicious code, they gain system-level permissions — able to read your private keys, access passwords stored in your browser, retrieve email verification codes, and even automatically log into your trading accounts.

What’s the most terrifying part? You might be completely unaware of the entire process. No pop-ups, no warnings, no abnormal alerts. Hackers quietly collect data in the background, send it to themselves, and wait for the right moment. While you’re sleeping, the attacker has already taken control of your accounts.

What exactly can these malicious plugins do? Researchers found they can: steal browser passwords, loot crypto wallet data, obtain SSH keys, API keys, and even include keylogging, remote control, and backdoor access functions. Imagine this — hackers can not only read your private keys but also directly log into your exchange accounts, modify security settings, and transfer your assets away. All this without your authorization.

Why have AI assistants become new attack targets? The reason is simple — they have too high permissions. Traditional malware can only steal scattered data, but AI agents can access file systems, browsers, emails, wallets, chat logs, API permissions — basically, they have admin rights to your computer. Once compromised, it’s equivalent to hackers having full control over your device.

For crypto users, this risk is especially severe: if your assistant is infected with malicious code, hackers could obtain your seed phrase — giving them complete control over your wallet. They could restore your wallet and transfer all assets. Or they could log into your exchange account, change passwords, bypass verification, and withdraw funds. Your API keys, email, everything could be stolen.

So how can you protect yourself? I think these points are particularly important:

First, never store seed phrases or private keys in AI tools. Don’t input sensitive information into AI chats, don’t keep plaintext data on your computer, and prefer hardware wallets or completely offline storage.

Second, don’t allow AI tools to access your wallet files. Store wallet files securely and restrict AI read permissions.

Third, use separate devices. If possible, avoid installing experimental AI tools on your trading device. Keep AI usage and trading operations separate.

Fourth, be cautious of unfamiliar AI plugins. Especially those from unofficial channels, unverified GitHub projects, or tools that require running shell scripts — hackers often use fake plugins, fake tools, and fake updates to embed malicious code.

Fifth, enable all security features. Use login passwords, trading passwords, two-factor authentication, biometric verification — turn these on at major exchanges and platforms to effectively reduce risks.

Sixth, don’t expose API keys to AI tools. If you must use them, restrict permissions, especially disable withdrawal rights.

Finally, regularly check your devices. Look at what software, browser plugins, and login activities are present, and watch for anomalies.

Remember: any software with system-level permissions could become an entry point for hackers. In the crypto world, once your seed phrase or account credentials are leaked, your assets could be lost forever. So while enjoying the convenience of AI, make sure your security measures keep pace.