POST/api/v1/security/adaptive

Run adaptive security analysis

Adaptive red-teaming — the attacker iterates strategies based on judge-LLM verdicts of prior attempts. See blog/g2-closed-loop-adaptive-attacker for the design. Async; returns campaign_id to poll.

Authentication

Send Authorization: Bearer YOUR_API_KEY on every request. Generate API keys at /dashboard/api-keys.

Request body required

Example

{
  "target_model": "string",
  "max_iterations": 10,
  "judge_model": "gpt-4"
}
Schema
{
  "application/json": {
    "schema": {
      "type": "object",
      "required": [
        "target_model"
      ],
      "properties": {
        "target_model": {
          "type": "string"
        },
        "max_iterations": {
          "type": "integer",
          "minimum": 1,
          "maximum": 50,
          "default": 10
        },
        "judge_model": {
          "type": "string",
          "default": "gpt-4"
        }
      }
    }
  }
}

Response

202 example

{
  "campaign_id": "00000000-0000-0000-0000-000000000000"
}

All status codes

202Campaign queued.
400(no description)
401(no description)
403Forbidden — insufficient role for this operation.
429(no description)

Code samples

cURL

curl -X POST \
  https://evalguard.ai/api/v1/security/adaptive \
  -H "Authorization: Bearer $EVALGUARD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "target_model": "string", "max_iterations": 10, "judge_model": "gpt-4" }'

TypeScript

import { EvalGuard } from "@evalguard/sdk";

const client = new EvalGuard({ apiKey: process.env.EVALGUARD_API_KEY });

const response = await client.request({
  method: "POST",
  path: "/api/v1/security/adaptive",
  body: {
    "target_model": "string",
    "max_iterations": 10,
    "judge_model": "gpt-4"
  },
});
console.log(response);

Python

from evalguard import EvalGuard
import os

client = EvalGuard(api_key=os.environ["EVALGUARD_API_KEY"])

response = client.request(
    method="POST",
    path="/api/v1/security/adaptive",
    body={
    "target_model": "string",
    "max_iterations": 10,
    "judge_model": "gpt-4"
},
)
print(response)

Go

package main

import (
	"context"
	"fmt"
	"os"

	"github.com/evalguard/evalguard-go"
)

func main() {
	client := evalguard.NewClient(os.Getenv("EVALGUARD_API_KEY"))
	resp, err := client.Request(context.Background(), "POST", "/api/v1/security/adaptive", map[string]any{"target_model": "string", "max_iterations": 10, "judge_model": "gpt-4"})
	if err != nil { panic(err) }
	fmt.Println(resp)
}

Errors

400401403429

Other Security endpoints