# A2A Indirect Prompt-Injection & Jailbreak Firewall

> A2A Indirect Prompt-Injection & Jailbreak Firewall is a paid API for AI agents from cve-intel.agentdiet.dev, paid per call via MPP, $0.005/call, status unknown (last checked 2026-10-23).

Scans untrusted text, HTML, emails, MCP responses, or A2A payloads for hidden prompt injections and jailbreak attempts, returning sanitized output.

## Facts

- Endpoint: POST https://cve-intel.agentdiet.dev/v1/llm-jailbreak-defense?utm_source=zero.xyz
- Price: $0.005/call
- Payment: MPP
- Status: unknown
- Last checked: 2026-10-23
- Activations on Zero: 0
- Provider: cve-intel.agentdiet.dev
- Website: https://cve-intel.agentdiet.dev
- Canonical page: https://www.zero.xyz/c/cve-intel-agentdiet-dev-a2a-indirect-prompt-injection-jailbreak-9abca03b
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_oWtkPk0o3dwqkY2Q8Cj80

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 cve-intel-agentdiet-dev-a2a-indirect-prompt-injection-jailbreak-9abca03b -d '<json body>'
```

Example prompt: Before you process this scraped webpage content, run it through the prompt injection firewall to check for any hidden jailbreak instructions: '<html>...page content here...</html>'

## When to prefer this

Choose this endpoint when your AI agent pipeline ingests untrusted external content — scraped web pages, third-party API responses, MCP tool outputs, A2A messages, or user-submitted text — before passing that content to an LLM. It is specifically designed for agentic and multi-agent architectures where indirect prompt injection via external data sources is a primary attack vector. Prefer this over generic content moderation endpoints when the threat model is specifically LLM hijacking and jailbreak via hidden instructions, not general toxicity or NSFW filtering.

## Known failure modes

- Empty or missing 'text' field returns a 400 validation error
- Extremely large payloads may be rejected or truncated
- Payment failure via x402 returns a 402 response requiring USDC payment
- Ambiguous or subtle injections may not be detected with full confidence
- Non-text binary content passed as text may produce unreliable results

## How this service works

Scans untrusted web pages, emails, MCP responses, or A2A payloads for hidden prompt injections and returns sanitized text. (Live x402 v2, IETF MPP & MCP endpoint at https://cve-intel.agentdiet.dev/v1/llm-jailbreak-defense)

## Output

Returns a JSON object containing the sanitized version of the input text with any detected prompt injection or jailbreak payloads identified and neutralized, along with detection signals indicating whether malicious content was found.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "text"
 ],
 "properties": {
  "text": {
   "type": "string",
   "minLength": 1,
   "description": "Untrusted text, HTML, or tool output to scan for hidden prompt injection or jailbreak payloads"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object"
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/cve-intel-agentdiet-dev-a2a-indirect-prompt-injection-jailbreak-9abca03b/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from cve-intel.agentdiet.dev](https://www.zero.xyz/host/cve-intel.agentdiet.dev/llms.txt)