Analyzes text to detect and flag prompt-injection attempts, helping secure AI prompts and agent interactions.