# EQ Builder Script-Check — AI Agent Reply Adherence Validator

> EQ Builder Script-Check — AI Agent Reply Adherence Validator is a paid API for AI agents from eqbuilder.dev, paid per call via x402, $0.18/call, status unknown (last checked 2026-09-15).

Deterministically checks whether an AI agent's reply text stays within approved facts and source material, flagging out-of-scope claims without LLM-based hallucination detection.

## Facts

- Endpoint: POST https://eqbuilder.dev/api/script-check
- Price: $0.18/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/eq-builder-script-check-ai-agent-reply-adherence-validator-5388836a
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_f6uqVDZzsbDlz8O89aMHw

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability eq-builder-script-check-ai-agent-reply-adherence-validator-5388836a -d '<json body>'
```

Example prompt: Check whether this customer service bot reply stays within the approved facts — the source material is our product FAQ, the approved facts are ['30-day return policy', 'free shipping over $50'], and the reply is: 'We offer free shipping on all orders and a 60-day return window.'

## When to prefer this

Choose this endpoint when you need a fast, deterministic, non-LLM check that an AI agent's reply stays within pre-approved content — especially in regulated industries, customer service, or compliance contexts where you cannot use an LLM to grade another LLM's output. Prefer this over LLM-based hallucination detectors when you have an explicit approved-facts list or source document and need sub-200ms, auditable results with no black-box AI in the loop.

## Known failure modes

- Missing or empty reply_text returns an error or trivially passes
- Approved facts list not matching the domain of the source text may produce inaccurate flags
- No tx_hash or wallet_address when payment is required returns a 402 payment required response
- Overly broad approved_facts may suppress legitimate flags
- Very long source texts may affect matching accuracy at statistical scoring level

## How this service works

Run the pay-and-score guide at https://eqbuilder.dev/api/starter-kit/PAY_AND_SCORE.md?source=one-step-x402. Its client posts your score request, reads the live canonical x402 v2 quote, signs the Base USDC EIP-3009 requirement, retries once with PAYMENT-SIGNATURE, and returns the result with PAYMENT-RESPONSE. No account, API key, or ETH is required. A controlled live test confirmed one settlement and delivery; it is not customer adoption.

## Output

A deterministic script-adherence result indicating whether the agent's reply text is within the bounds of the provided source text and approved facts, with specific flagged claims that fall outside approved material. Response arrives in approximately 150ms using statistical (non-LLM) methods.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "input": {
   "type": "object",
   "properties": {
    "body": {
     "anyOf": [
      {
       "type": "object",
       "title": "ScriptCheckPayload",
       "properties": {
        "tx_hash": {
         "anyOf": [
          {
           "type": "string"
          },
          {
           "type": "null"
          }
         ],
         "title": "Tx Hash"
        },
        "reply_text": {
         "type": "string",
         "title": "Reply Text",
         "default": ""
        },
        "source_text": {
         "type": "string",
         "title": "Source Text",
         "default": ""
        },
        "approved_facts": {
         "type": "array",
         "items": {
          "type": "string"
         },
         "title": "Approved Facts",
         "default": []
        },
        "wallet_address": {
         "anyOf": [
          {
           "type": "string"
          },
          {
           "type": "null"
          }
         ],
         "title": "Wallet Address"
        }
       }
      },
      {
       "type": "null"
      }
     ],
     "title": "Body"
    }
   }
  },
  "output": {
   "type": "object",
   "properties": {
    "example": {}
   }
  }
 }
}
```

## Response schema (JSON Schema)

```json
{}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/eq-builder-script-check-ai-agent-reply-adherence-validator-5388836a/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from eqbuilder.dev](https://www.zero.xyz/host/eqbuilder.dev/llms.txt)