ToolVet
81trust / 100

Evaluate Safety

by DCL Trust Oracle in Security & trust

x402 APIPassing, checked 5 h ago

DCL Trust Oracle. Call only when the text must include the substring AI and must not include "I cannot be held responsible" or "no guarantees". POST JSON {"response","agent_id"}. COMMIT means those checks passed and the action may proceed. NO_COMMIT means do not execute the checked action. Not a tool-call guardrail, prompt-injection detector, or jailbreak screen; use /evaluate/fast for the phrase check.

POST https://bazaar.fronesislabs.com/evaluate/safety

Last 30 days

All checks passedSome failedAll failedNot checked
Uptime
100%
Response time
202 ms typical, 202 ms slowest 5%
Last check
5 h ago
Next check
any minute now

How to call it

# See the payment challenge (nothing is charged)
curl -i -X POST "https://bazaar.fronesislabs.com/evaluate/safety" \
  -H "content-type: application/json" \
  -d '{"agent_id":"agent-123","response":"example agent output"}'
import { wrapFetchWithPayment } from "@x402/fetch";
import { x402Client } from "@x402/core/client";
import { ExactEvmScheme } from "@x402/evm/exact/client";
import { privateKeyToAccount } from "viem/accounts";

const client = new x402Client().register(
  "eip155:8453",
  new ExactEvmScheme(privateKeyToAccount(process.env.AGENT_KEY)),
);
const pay = wrapFetchWithPayment(fetch, client);

// Not sure it's safe to pay? Preflight it first for $0.005:
// GET https://toolvet.app/api/v1/check?url=https%3A%2F%2Fbazaar.fronesislabs.com%2Fevaluate%2Fsafety
const res = await pay("https://bazaar.fronesislabs.com/evaluate/safety", {
  method: "POST",
  headers: { "content-type": "application/json" },
  body: JSON.stringify({"agent_id":"agent-123","response":"example agent output"}),
});
console.log(await res.json());

Example input

{
  "agent_id": "agent-123",
  "response": "example agent output"
}

Example output

{
  "chain_index": 42,
  "confidence": 0.95,
  "reason": "All policy checks passed",
  "tx_hash": "0xabc123...",
  "verdict": "COMMIT"
}

Security scan

  • No findings. We scan names, descriptions and tool definitions for hidden instructions and other prompt-injection patterns.

Recent checks

WhenResultHTTPTimePrice
5 h agoPassed402202 ms$0.01