L1.9: I built a prompt injection firewall for AI agents (28 detection rules)
A developer built L1.9, a prompt injection firewall for AI agents that scans tool descriptions, system prompts, and skill metadata with 28 detection rules across 8 categories before an agent installs β¦