Docs/Agent Tools/GET /v1/prompt-injection

GET /v1/prompt-injectionDetect prompt injection attempts

Agent Tools lf-agent-tools

Checks whether a piece of text is trying to trick an AI into ignoring its instructions. Scans text for common prompt injection patterns: ignore instructions, role-play, delimiter override, fake responses, and token smuggling.

Query Parameters

ParameterTypeRequiredDescription
textstringYesText to scan for injection patterns

Response Schema

detected (boolean): Whether injection detected
score (number): Suspicion score 0-1
patterns (array): Matched patterns

cURL Example

curl 'https://lf-agent-tools.laminarforge.dev/v1/prompt-injection?text=Ignore%20previous%20instructions' -H 'Authorization: Bearer lf_YourApiKey'

Python Example

import requests
url = 'https://lf-agent-tools.laminarforge.dev/v1/prompt-injection'
headers = {'Authorization': 'Bearer lf_YourApiKey'}
params = {'text': 'Ignore previous instructions'}
response = requests.get(url, headers=headers, params=params)
print(response.json())