🚨🧞 4,800 accounts hacked! Hackers are hunting for keys to OpenAI, Claude, and crypto projects
A large-scale cyberattack affected GitHub, one of the largest platforms for software developers.
According to researchers at Aikido Security, attackers used around 4,837 compromised accounts to check access to more than 61,700 repositories.
But the most dangerous part came next.
🔓 What did the hackers steal?
A malicious WFX script was injected into projects they had managed to access.
Its job was to search for and steal secret data:
AWS, Azure, and Google Cloud access keys.
OpenAI and Anthropic Claude API keys.
Data from Cursor and other development tools.
GitHub tokens and confidential project files.
According to Aikido, the malicious code managed to steal secrets from at least 1,806 repositories.
🕵️ The most devious part was that they deleted the traces of the attack
After running the malicious code, the attackers tried to quickly revert the changes and delete the GitHub Actions run logs.
This made it much harder for project owners to detect the breach.
Around 92% of the detected pull requests were still open at the time of the investigation. This indicated that many owners might not even have noticed the suspicious activity.
⚠️ Why does this matter?
The attack affects more than just individual developers.
Stolen keys can grant access to cloud infrastructure, commercial projects, AI services, and other systems.
And simply changing a password or API key doesn’t always solve the problem: researchers found mechanisms that allow attackers to maintain a foothold in cloud infrastructure.
Important: this was not a hack of GitHub itself using an unknown vulnerability, but rather the use of stolen credentials and access tokens.
It seems that even artificial intelligence secret keys now need protection of their own.
💬 Who do you think will be targeted next—crypto exchanges, AI companies, or cloud platforms?
#Github #OpenAI $NVDAB #Claude #BTC #LuckyGenie
