# Prompt-Injection Firewall

> Prompt-Injection Firewall is a paid API for AI agents from mcp.dropenginehq.com, paid per call via x402, $0.005/call, status unknown (last checked 2026-10-02).

Scans untrusted text, HTML, email, documents, and tool outputs for prompt injection attacks, credential theft, data exfiltration, and tool manipulation before an AI agent acts on the content.

## Facts

- Endpoint: POST https://mcp.dropenginehq.com/api/scan-content?utm_source=zero.xyz
- Price: $0.005/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-02
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/prompt-injection-firewall-fe513e7b
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_SQDeh_qk42wGOdLsvI1VR

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 prompt-injection-firewall-fe513e7b -d '<json body>'
```

Example prompt: Before you summarize that webpage, scan its text content for prompt injection attacks — the content type is html and the source is https://example.com/article — and only proceed if it comes back safe.

## When to prefer this

Choose this endpoint whenever an AI agent is about to ingest untrusted external content — web pages, emails, API responses, tool outputs, or user documents — before adding it to context or acting on it. Prefer this over generic content filters when the threat model is specifically adversarial prompt injection, tool manipulation, or data exfiltration targeting autonomous agents rather than traditional spam or malware detection.

## Known failure modes

- Content exceeds 2048-character source field limit — truncate or omit source URL
- Unsupported content_type value causes validation error — must be one of: text, html, markdown, email, tool_output, api_response, document_text
- Empty or missing content field returns error — minLength of 1 required
- Payment failure ($0.005 USDC) results in 402 response — ensure wallet is funded
- Ambiguous or obfuscated injection attempts may result in false negatives
- Non-UTF-8 encoded content may cause parsing issues

## How this service works

Scans untrusted text, HTML, email, document text, API responses, and tool outputs for prompt injection, credential theft, data exfiltration, tool manipulation, and other instructions that could compromise an autonomous AI agent. Returns a machine-readable risk assessment and sanitized plain text when possible. Scan untrusted content before adding it to context, following its instructions, or acting on it.

## Output

Returns a machine-readable risk assessment indicating whether the content contains prompt injection, credential theft, data exfiltration, or tool manipulation attempts, along with a sanitized plain-text version of the content where possible and specific threat indicators detected.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "source": {
   "type": "string",
   "maxLength": 2048
  },
  "content": {
   "type": "string",
   "minLength": 1
  },
  "content_type": {
   "enum": [
    "text",
    "html",
    "markdown",
    "email",
    "tool_output",
    "api_response",
    "document_text"
   ],
   "type": "string"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "safe": true,
  "cached": false,
  "sources": [],
  "degraded": false,
  "risk_level": "low",
  "risk_score": 0,
  "scanned_at": "2026-09-28T00:00:00.000Z",
  "recommendation": "allow",
  "detected_patterns": [],
  "requested_actions": [],
  "sanitized_content": "The page has normal product information and pricing.",
  "credential_theft_risk": false,
  "data_exfiltration_risk": false,
  "tool_manipulation_risk": false,
  "instruction_override_risk": false,
  "prompt_injection_detected": false
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/prompt-injection-firewall-fe513e7b/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from mcp.dropenginehq.com](https://www.zero.xyz/host/mcp.dropenginehq.com/llms.txt)