My blog was hacked and Claude and I just fixed it (lengrand.fr)

A tech blogger recently shared their experience of discovering and fixing a security breach on their self-hosted Ghost blog after receiving reports of fake captchas. Initially skeptical about the claims, they conducted an investigation using Claude Code, an AI tool, which helped identify the root of the problem: a malicious script injected into the database of their blog posts, specifically targeting Windows users. The incident underscores the importance of vigilance in website security, particularly concerning third-party software. Through thorough investigation and database queries with SQL, the blogger was able to nullify the harmful code injections and implement preventive measures, including server updates and password changes. This experience highlights the dual role of AI as both a tool for assistance and a reminder of the potential pitfalls of reliance on technology, urging professionals in the AI/ML community to remain cautious and proactive about cybersecurity.