Post

🤔 Should you use AI to investigate security alerts? The one investigating a security alert is not usually who wrote the code that triggered the alert, so they struggle investigating if it’s a real security incident. 🤖 Can AI help here? Check the full experiment at: www.sysdig.com/blog/investi...

lnkd.inLinkedInThis link will take you to a page that’s not on LinkedIn

14/11/25, 4:54 pm0 reposts0 likes

RootNode.es

@rootnode.es

14 Nov 2025

This article covers this use case by connecting Codex to several MCPs. First, retrieves a list of alerts using the Sysdig MCP Then investigates the issue with the GitHub MCP without downloading the code. It ends by creating a GitHub issue with context for the relevant team to implement a fix.

RootNode.es

@rootnode.es

14 Nov 2025

🔥 Codex can help with the basics. It correctly identifies the source of the alert and can create a GitHub issue. 💣 However, it doesn’t explain things correctly and can be manipulated to ignore some data. You must supervise it.