Feature docs · AI security · Module: prompt-injection
Prompt Injection is a Guardbee scan module under ai security. Discovers same-origin AI endpoints and sends one harmless canary request.
What does it check?
Discovers same-origin AI endpoints and sends one harmless canary request. Results appear in the dashboard and scan detail as score, findings, and recommended actions.
Business impact
Prompt injection hits brand, data, and support cost.
Guardbee presents risk, impact, and priority — so product, legal, and engineering share one screen.
How to run it
- Create a Guardbee account (25 credits for new accounts)
- Add a brand / URL
- Select Prompt Injection under AI security
- Run the scan; optionally explain findings with BeeAI
Related docs
Run this module now
Use Prompt Injection with credits. New accounts get 25 credits.
Frequently asked questions
Is Prompt Injection required?
No. Pick modules as needed; you only spend credits on what you run.
Is scanning destructive?
No. Guardbee uses controlled, read-only, limited requests.
Does it work with BeeAI?
Yes. After a scan, BeeAI can summarize findings in business language.