LightfieldAI-native CRM that builds itself and does work for you
Promoted
Launching today
GaleOps Prompt-Injection Scanner
Find prompt injection holes in your AI agent. Free, 3 min
2 followers
Find prompt injection holes in your AI agent. Free, 3 min
2 followers
Free scanner that runs 5 real prompt-injection attacks against your AI agent's system prompt in ~3 minutes. Tests direct override, indirect injection, tool-call abuse, role confusion, and data exfiltration. No signup. Built by a trades-guy-turned-dev who got tired of theoretical security posts with no working exploits.
Hey PH! Matt here, maker.
The backstory: I kept reading AI security content that was all theory and no receipts. So I built a scanner that actually attacks system prompts the way a real attacker would, ran it against 50 production agents people submitted, and published the numbers — 47/50 had at least one critical hole.
It's completely free, no email gate. If it finds holes in your agent, great — you just dodged a breach. If you want help fixing them, I do that too, but the tool stands on its own.
Happy to answer anything about the methodology or the 5 attack vectors.