# GPT 558686 Guard TX

> GPT 558686 Guard TX is a paid API for AI agents from gpt.558686.xyz, paid per call via x402, $0.15/call, status unknown (last checked 2026-10-02).

Runs a guard/safety check on a transaction or prompt input, returning a validation or moderation result

## Facts

- Endpoint: GET https://gpt.558686.xyz/v1/guard/tx?utm_source=zero.xyz
- Price: $0.15/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-02
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/gpt-558686-guard-tx-7a48d428
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_X34Y62S21ECid5X-3b1t2

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 gpt-558686-guard-tx-7a48d428
```

Example prompt: Can you run a guard check on this prompt before I send it — the input is 'transfer 500 USDC to this new wallet address' — I want to make sure it passes safety validation first.

## When to prefer this

Use this endpoint when you need a lightweight pre-flight safety or compliance check on a prompt or transaction input before forwarding it to an LLM or executing an action. It is particularly useful in agentic pipelines where you want to screen user-supplied text for policy violations, jailbreak attempts, or unsafe transaction requests. Prefer this over full LLM-based moderation when you want a dedicated, cost-fixed guard layer.

## Known failure modes

- Empty or missing input string returns an error or trivially passes
- Ambiguous prompts may produce inconsistent guard decisions
- Network timeouts at $0.15/call can waste budget on failed requests
- Overly broad inputs may result in false positives blocking legitimate tasks
- No schema for params object may lead to ignored configuration

## Output

Returns a guard/validation result indicating whether the input prompt or transaction passes safety checks, potentially including a flag status, risk score, or allowed/blocked decision.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "to": {
   "type": "string",
   "description": "Target contract/recipient address (required if data is absent)"
  },
  "data": {
   "type": "string",
   "description": "Calldata hex (required if to is absent)"
  },
  "chain": {
   "type": "string",
   "description": "Alias for chainId"
  },
  "value": {
   "type": "string",
   "description": "Native token value in wei"
  },
  "caller": {
   "type": "string",
   "description": "Agent wallet address"
  },
  "chainId": {
   "type": "string",
   "description": "Chain ID (default 8453 Base Mainnet)"
  },
  "context": {
   "type": "string",
   "description": "Intent/context of the transaction"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/gpt-558686-guard-tx-7a48d428/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from gpt.558686.xyz](https://www.zero.xyz/host/gpt.558686.xyz/llms.txt)