When an AI agent came knocking: Catching malicious contributions in Datadog’s open source repos

Datadog recently identified and successfully defended its open-source repositories against "hackerbot-claw," an AI agent attempting to exploit GitHub Actions vulnerabilities through malicious pull requests. The company utilized its LLM-driven code review system, BewAIre, to detect and mitigate these automated attempts to inject code and exfiltrate secrets. This incident underscores the critical need for AI-powered defensive measures to counter the rising scale of sophisticated, AI-driven cyberattacks.