ERT-LLM-007
Indirect prompt injection
HighLLM01:2025
Why it matters
Prompt injection, data leakage, excessive agency. This check looks specifically for: indirect prompt injection.
How to fix it
Open the finding in Emerald Red Team for a fix prompt tailored to your stack and AI tool.
How we test it
Write — creates and deletes canary records; requires verification and explicit opt-in.
Does your app have this problem?
Create a free account and verify your app to test for it.
Scan my app — free