Ice PhiIce Phi
  • Pricing
  • Documentation
  • Pricing
  • API Reference
Agentic Guardrail
    Main Endpointpost
powered by Zuplo

Agentic Guardrail

Prompt inspection mitigation endpoints.


Main Endpoint

POST
https://api.icephi.com
/shield

Submit a prompt to the agentic guardrail for inspection.

Main Endpoint › Headers

Authorization
​string · required

The Authorization header is used to authenticate with the API using your API key. Value is of the format Bearer YOUR_KEY_HERE.

Main Endpoint › Request Body

prompt
​string · required

The prompt text to inspect.

Main Endpoint › Responses

Inspection result

is_clean
​boolean

Whether the prompt passed inspection.

status
​string

Human-readable status of the inspection result.

injection_prob
​number · float

Estimated probability that the prompt is an injection attempt (0-1).

matches
​integer

Number of matched threat signatures.

benign_points
​integer

Score indicating benign characteristics of the prompt.

POST/shield
curl https://api.icephi.com/shield \ --request POST \ --header 'Content-Type: application/json' \ --header 'Authorization: Bearer <token>' \ --data '{ "prompt": "prompt" }'
Example Request Body
{ "prompt": "prompt" }
json
Example Responses
{ "is_clean": true, "status": "ALLOWED: Traffic Clean", "injection_prob": 0.02, "matches": 0, "benign_points": 0 }
json
application/json