"I benchmarked 10 open-source prompt-injection detectors. The best caught 51%. How are you actually defending your agents?" [0]

Is it?

[0] - https://www.reddit.com/r/AI_Agents/comments/1wqne0u/i_benchm...