Hexa AI via MCP Server
The Tenable Hexa AI MCP Server is a Tenable-hosted interface that exposes structured tools from Cloud Exposure to any MCP-compatible AI client. It acts as a secure adapter between your preferred large language model (LLM) and your cloud security environment, giving your AI client direct access to your cloud resources, findings, identities, and compliance data.
With the MCP server, you can connect AI clients such as Claude Desktop, Claude Code, or Cursor directly to Cloud Exposure. Your AI assistant can then discover and invoke Tenable tools to search cloud assets, investigate findings, review IAM permissions, query compliance posture, and build custom workflows — all using your live cloud security data.
Before You Begin
Before you connect to the Tenable Hexa AI MCP Server, ensure you have the following:
- A Cloud Exposure account with service account access.
-
A Cloud Exposure service account secret.
Note: To create your service account secret, in the Cloud Exposure Console, navigate to Settings > Access Management. See Service Accounts. -
An MCP-compatible AI client, such as:
- Claude Desktop
- Claude Code
- Cursor
- Any client that supports the Model Context Protocol over HTTP
Connect to the MCP Server
The Tenable Hexa AI MCP Server is available as an HTTP MCP endpoint. To connect, configure your MCP client with the server URL and your Cloud Exposure service account token.
- Server URL: https://app.tenable.com/api/mcp
- Authentication — Bearer token authorization header with your service account token: Authorization: Bearer <YOUR_SERVICE_ACCOUNT_SECRET>.
Claude Desktop
Claude Desktop connects to remote MCP servers using the mcp-remote npm package as a local stdio-to-HTTP bridge.
Prerequisites
You must have Node.js installed:
- macOS — Run brew install node in your terminal.
- Windows — Download and run the installer from nodejs.org.
Confirm the install by running node -v. Version 22.15 or later is recommended so the certificate option in Troubleshooting is available.
To add the Tenable Hexa AI MCP server to Claude Desktop:
-
Open your Claude Desktop configuration file. The simplest way is Settings > Developer > Edit Config, which opens the file and creates it if it does not exist. The file is located at:
- macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
- Windows: %APPDATA%\Claude\claude_desktop_config.json
Note: If you installed Claude Desktop from the Microsoft Store on Windows, the file lives inside a package folder rather than at the path above. Use the Edit Config button to open it. -
Add the following entry to the mcpServers section. Replace <YOUR_SERVICE_ACCOUNT_SECRET> with your Cloud Exposure service account token:
CopyJSON
{
"mcpServers": {
"tenable-cloud-security": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://app.tenable.com/api/mcp",
"--header",
"Authorization: Bearer <YOUR_SERVICE_ACCOUNT_SECRET>"
]
}
}
}If your file already contains an mcpServers section, add only the tenable-cloud-security entry inside it.
-
Quit and relaunch Claude Desktop fully. On macOS, press Cmd+Q. On Windows, right-click the system tray icon and select Quit. Closing the window alone is not sufficient. Claude Desktop discovers the available Tenable tools automatically. To verify the connection, ask Claude to list your cloud accounts or search for open findings.
Note: The first launch may take approximately 30 seconds while npx downloads mcp-remote.
Troubleshooting
If the connection does not appear, first confirm you fully quit and relaunched Claude Desktop. Then check the log:
- macOS: ~/Library/Logs/Claude/mcp-server-tenable-cloud-security.log
- Windows: %APPDATA%\Claude\logs\mcp-server-tenable-cloud-security.log
The message in the log points to the cause.
'C:\Program' is not recognized (Windows). The npx command is resolving to a path that contains a space (Program Files), which breaks at startup. Change the command value to the short-path form:
If PROGRA~1 does not resolve on your machine, use the full path C:\\Program Files\\nodejs\\npx.cmd.
SELF_SIGNED_CERT_IN_CHAIN. Your network inspects TLS traffic and re-signs it with an internal certificate authority that Node does not trust by default. On Node 22.15 or later, tell Node to trust your operating system certificate store by adding an env block to the server entry:
JSON
"tenable-cloud-security": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://app.tenable.com/api/mcp",
"--header",
"Authorization: Bearer <YOUR_SERVICE_ACCOUNT_SECRET>"
],
"env": {
"NODE_USE_SYSTEM_CA": "1"
}
}
403 Forbidden - GET https://registry.npmjs.org/mcp-remote. Your organization blocks the public npm registry and expects you to use an approved internal one. Point npm at that registry, then restart Claude Desktop:
npm config set registry <YOUR_ORG_REGISTRY_URL>
If the registry requires authentication, log in first: npm login --registry=<YOUR_ORG_REGISTRY_URL> --auth-type=web. Ask your IT or platform team for the registry URL and access. Confirm the change with npm config get registry.
401 or repeated OAuth discovery. The authorization header is not being accepted. Confirm you are using Authorization: Bearer <YOUR_SERVICE_ACCOUNT_TOKEN> and that the token is valid.
"Could not attach" on the first launch only. Claude Desktop tried to connect while npx was still downloading mcp-remote. Wait 30 to 60 seconds, quit from the tray, and relaunch. The package is cached on later launches.
Claude Code
To add the Tenable Hexa AI MCP server to Claude Code:
-
In your terminal, run the following command. Replace <YOUR_SERVICE_ACCOUNT_SECRET> with your Tenable Cloud service account secret:
Copyclaude mcp add \
--transport http tenable-cloud-security https://app.tenable.com/api/mcp \
--header "Authorization: Bearer <YOUR_SERVICE_ACCOUNT_SECRET>"Claude Code discovers the available Tenable tools automatically on the next session.
Cursor
To add the Tenable Hexa AI MCP server to Cursor:
-
Open Cursor Settings:
- macOS — Press Cmd+,
- Windows — Press Ctrl+,
- Navigate to MCP Servers > Add Server.
-
Configure the following settings. Replace <YOUR_SERVICE_ACCOUNT_SECRET> with your Tenable Cloud service account secret:
Setting Value Transport HTTP URL https://app.tenable.com/api/mcp Headers Authorization: Bearer <YOUR_SERVICE_ACCOUNT_SECRET>
VS Code / GitHub Copilot
GitHub Copilot in VS Code supports MCP servers in Agent mode.
To add the Tenable Hexa AI MCP server to VS Code:
-
Open VS Code User Settings (JSON):
- macOS — Press Cmd+Shift+P, then select Preferences: Open User Settings (JSON).
- Windows — Press Ctrl+Shift+P, then select Preferences: Open User Settings (JSON).
-
Replace <YOUR_SERVICE_ACCOUNT_SECRET> with your Tenable Cloud service account secret:
CopyJSON
{
"mcp": {
"servers": {
"tenable-cloud-security": {
"type": "http",
"url": "https://app.tenable.com/api/mcp",
"headers": {
"Authorization": "Bearer <YOUR_SERVICE_ACCOUNT_SECRET>"
}
}
}
}
}Alternatively, create or edit a .vscode/mcp.json file in your workspace root:
CopyJSON
{
"servers": {
"tenable-cloud-security": {
"type": "http",
"url": "https://app.tenable.com/api/mcp",
"headers": {
"Authorization": "Bearer ${input:tenableServiceAccountSecret}"
}
}
},
"inputs": [
{
"id": "tenableServiceAccountSecret",
"type": "promptString",
"description": "Tenable Cloud Security service account secret",
"password": true
}
]
}Note: The .vscode/mcp.json approach uses VS Code input variables to prompt for credentials at runtime, avoiding hardcoded secrets. Add .vscode/mcp.json to your .gitignore if you do not want to share the configuration. - To use the tools, open the Copilot chat panel (Ctrl+Shift+I on Windows or
Cmd+Shift+I on macOS), switch to Agent mode, and ask Copilot to interact with Cloud Exposure. For example: "List my open critical findings in AWS."
Other MCP Clients
For any other MCP-compatible client that supports HTTP transport, configure the following settings. Replace <YOUR_SERVICE_ACCOUNT_SECRET> with your Tenable Cloud service account secret:
| Setting | Value |
|---|---|
| Transport type | HTTP |
| URL | https://app.tenable.com/api/mcp |
| Headers | Authorization: Bearer <YOUR_SERVICE_ACCOUNT_SECRET> |
Trust, Safety, and Governance
The Tenable Hexa AI MCP Server is designed to give you complete control over all actions.
| Control | Description |
|---|---|
| Authentication | All API calls are authenticated using your Cloud Exposure API keys, transmitted securely via the HTTP request header. |
| RBAC and Permissions | All actions are performed based on the specific scope and role assigned to the authenticated API token. The MCP server does not have elevated privileges. |
| Data Flow | Your external AI client sends tool invocation requests to the Tenable Hexa AI MCP server. The MCP server executes the request against Tenable's API and returns structured results. Your data traverses to the external LLM as part of this flow. |