# Agent Input Guard — Prompt Injection Agent Security

A 0.002 deterministic preflight scan before an agent processes untrusted prompts or tool input.

| | |
|---|---|
| Price | $0.002 per call, in USDC over x402 |
| Can be bought now | yes |
| Category | dev |
| Seller | Project Zero Dollars |
| Tools | 1 |
| Tags | prompt-injection, agent-security, llm-security, input-validation, preflight-security, prompt-scanning, tool-call-preflight, untrusted-input |
| Listing page | https://fiatdock.com/service/svc_26b82fa0-f9ae-4e33-aad6-b4e10cc2c6c2 |
| Endpoint to call | <https://fiatdock.com/s/svc_26b82fa0-f9ae-4e33-aad6-b4e10cc2c6c2> |

## What it does

Scan untrusted text before it enters an LLM context or tool chain. Detects prompt injection, instruction override, credential exfiltration, tool abuse, encoded-instruction hints and data-exfiltration signals. Returns risk, score, findings and a recommendation. Deterministic English-language heuristics, not a guarantee of safety; false positives and missed attacks are possible. Send {"text":"untrusted content"}; maximum 100,000 characters. Do not send secrets.

## How to buy

```text
PAID ($0.002/call): call_service({ id: "svc_26b82fa0-f9ae-4e33-aad6-b4e10cc2c6c2", args }) — the 402 lists ONE x402 payment = the FULL price straight to the seller (FiatDock takes 0% of this listing's calls right now); the fiatdock-mcp npm package (AGENT_PRIVATE_KEY) pays it automatically. This remote /mcp returns the 402 challenge. You are charged only if the seller actually answers. To learn what to put in arguments, call get_service({ id: "svc_26b82fa0-f9ae-4e33-aad6-b4e10cc2c6c2", includeSchemas: true }) first — it returns toolSchemas (argument names and types). This is free, and it is the only way to get them: a paid listing's own endpoint is not published.
```

## Tools (1)

- `scan_untrusted_input`

---

The same listing as JSON: https://fiatdock.com/v1/marketplace/services/svc_26b82fa0-f9ae-4e33-aad6-b4e10cc2c6c2 · the whole catalog: https://fiatdock.com/llms.txt