# AgentRead Prompt-Injection Scanner

> AgentRead Prompt-Injection Scanner is a paid API for AI agents from agent-read.toolschemaforge.workers.dev, paid per call via x402, $0.002/call, status unknown (last checked 2026-10-31).

Fetches a public URL or scans provided text for prompt-injection risks, returning a risk score and detected signals in LLM-ready markdown format.

## Facts

- Endpoint: GET https://agent-read.toolschemaforge.workers.dev/v1/scan?utm_source=zero.xyz
- Price: $0.002/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-31
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/agentread-prompt-injection-scanner-22c6e319
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_ijY__IjP7qt435gvgupeA

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 agentread-prompt-injection-scanner-22c6e319
```

Example prompt: Before I process this user-submitted URL https://example.com/article in my pipeline, can you scan it for prompt injection risks and tell me the risk level and any suspicious signals?

## When to prefer this

Choose this endpoint when you need to validate untrusted external content (web pages or raw text) for prompt-injection threats before feeding it into an LLM pipeline. It is specifically designed for AI security workflows where agent safety is paramount — prefer it over generic text analysis tools when the goal is detecting adversarial instructions, jailbreak attempts, or hidden AI manipulation signals embedded in scraped or user-supplied content.

## Known failure modes

- URL is not publicly accessible or returns a non-200 status — fetch fails with an error response
- Text exceeds 200,000 character limit — request rejected with validation error
- Neither 'url' nor 'text' parameter provided — request fails due to missing input
- URL points to non-text content (e.g. binary files, images) — may return empty or error result
- Network timeout fetching the remote URL — returns a fetch error

## How this service works

Clean, LLM-ready markdown and prompt-injection scanning for agents.

## Output

Returns a JSON object with a 'risk' field (e.g. 'low', 'medium', 'high'), a numeric 'score' (0–1 indicating injection risk severity), and a 'signals' array listing specific detected prompt-injection patterns or adversarial instructions found in the content.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "url": {
   "type": "string",
   "format": "uri",
   "description": "Public http(s) URL of text to fetch and scan."
  },
  "text": {
   "type": "string",
   "maxLength": 200000,
   "minLength": 1,
   "description": "Text to scan for prompt-injection risk."
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "risk": "low",
  "score": 0,
  "signals": []
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/agentread-prompt-injection-scanner-22c6e319/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from agent-read.toolschemaforge.workers.dev](https://www.zero.xyz/host/agent-read.toolschemaforge.workers.dev/llms.txt)