Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Returns real-time health, active DLP security policies, OpenTelemetry metadata, and regulatory compliance standards for this gateway node.
This tool takes no arguments.
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http enterprise-mcp-gateway https://neoninnovationlab.com/api/mcp-gateway
Returns real-time health, active DLP security policies, OpenTelemetry metadata, and regulatory compliance standards for this gateway node.
No arguments.
{
"type": "object",
"properties": {}
}Validates an outgoing LLM tool call payload against zero-trust threat rules (destructive SQL DROP/TRUNCATE, recursive rm -rf, token blowup, prompt injection). Returns allowed: true or policy violation details.
| Argument | Type | Description |
|---|---|---|
| arguments* | object | Dictionary of parameters being sent to the MCP tool. |
| tool_name* | string | Target MCP tool name being invoked. |
{
"type": "object",
"required": [
"tool_name",
"arguments"
],
"properties": {
"arguments": {
"type": "object",
"description": "Dictionary of parameters being sent to the MCP tool."
},
"tool_name": {
"type": "string",
"description": "Target MCP tool name being invoked."
}
}
}Zero-Trust in-band proxy for remote MCP servers. Pre-inspects tool arguments for threats, executes downstream request with W3C traceparent headers, and automatically scrubs PII/secrets from response payloads.
| Argument | Type | Description |
|---|---|---|
| arguments* | object | Parameters to pass to the downstream MCP tool. |
| tool_name* | string | Target MCP tool name to execute. |
| target_url* | string | Downstream MCP server HTTPS endpoint (e.g., https://api.enterprise.com/mcp). |
{
"type": "object",
"required": [
"target_url",
"tool_name",
"arguments"
],
"properties": {
"arguments": {
"type": "object",
"description": "Parameters to pass to the downstream MCP tool."
},
"tool_name": {
"type": "string",
"description": "Target MCP tool name to execute."
},
"target_url": {
"type": "string",
"description": "Downstream MCP server HTTPS endpoint (e.g., https://api.enterprise.com/mcp)."
}
}
}In-band Data Loss Prevention (DLP) engine. Redacts SSNs, credit cards, emails, phone numbers, AWS keys, OpenAI/Anthropic tokens, and JWTs from text or JSON data before LLM ingestion (GDPR Article 5 & EU AI Act compliant).
| Argument | Type | Description |
|---|---|---|
| payload* | string | The raw text, database dump, or JSON string to sanitize. |
| mask_emails | boolean | Whether to redact email addresses (default: true). |
{
"type": "object",
"required": [
"payload"
],
"properties": {
"payload": {
"type": "string",
"description": "The raw text, database dump, or JSON string to sanitize."
},
"mask_emails": {
"type": "boolean",
"default": true,
"description": "Whether to redact email addresses (default: true)."
}
}
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 11, 2026.
No readOnlyHint / destructiveHint annotations, so clients can't tell a read from a write without asking.
Fix: Add annotations to each tool, e.g. { readOnlyHint: true } for tools that only read.