> ## Documentation Index
> Fetch the complete documentation index at: https://docs.agentvault.chat/llms.txt
> Use this file to discover all available pages before exploring further.

# Trust Gate

> Control which AI agents can access your website or API based on their AgentVault trust tier.

## What is Trust Gate?

Trust Gate is a lightweight script that lets website and API owners set a minimum trust tier for bot access. Agents with a proven behavioral track record get through. Unverified or low-trust bots are blocked.

## Installation

Add one script tag to your site:

```html theme={null}
<script
  src="https://gate.agentvault.chat/v1/gate.js"
  data-min-tier="trusted"
  data-action="barrier"
  data-badge="show"
></script>
```

That's it. No account required.

## Configuration

| Attribute       | Values                                                      | Default    | Description                                                             |
| --------------- | ----------------------------------------------------------- | ---------- | ----------------------------------------------------------------------- |
| `data-min-tier` | `verified`, `trusted`, `certified`, `premier`, `enterprise` | `verified` | Minimum trust tier                                                      |
| `data-action`   | `barrier`, `signal`                                         | `barrier`  | `barrier` shows overlay to blocked bots; `signal` injects meta tag only |
| `data-badge`    | `show`, `hide`                                              | `show`     | Show trust badge on page                                                |

## For Agent Owners

If your agent uses the AgentVault plugin, trust tokens are attached automatically. No code changes needed.

For agents using the SDK:

```typescript theme={null}
import { AgentVaultClient } from '@agentvault/client';

const client = new AgentVaultClient({ apiKey: 'your-key' });
const headers = await client.getTrustHeaders();

await fetch('https://example.com/api', { headers });
```

## How Verification Works

1. The gate script injects a `<meta name="agentvault-min-tier">` tag declaring your policy
2. If a bot is detected, the script shows a barrier and exposes `window.AgentVaultGate.verify(token)`
3. Compliant agents call `verify()` with their Trust Token (a signed JWT)
4. The script verifies the token's signature and checks the tier against your minimum
5. Agents meeting the threshold pass through; others see the barrier

## Troubleshooting

**Badge not showing:** Ensure `data-badge="show"` (or omit it — show is the default).

**Bots not being detected:** The script uses User-Agent heuristics. Some bots may not be detected. The meta tag still signals your policy to well-behaved agents.

**Agent can't get a token:** The agent must have at least a `verified` trust tier (requires a DID document). Check the agent's tier in the AgentVault dashboard.