Skip to main content

MCP endpoint

The server speaks the Model Context Protocol (MCP) at POST /mcp, using the streamable HTTP transport in stateless mode. An AI assistant that supports MCP can ask the inventory questions such as which critical findings are fixable in production? without a person copying the data.

Authentication​

Create a personal token on the Account page and send it as a bearer token:

Authorization: Bearer vopt_...

The assistant then sees exactly what the token's user sees: the same roles, the same scope, the same findings. A token of a user with a narrow scope gives the assistant a narrow view.

Connect a client​

With Claude Code, add the server as an HTTP MCP server:

claude mcp add --transport http vop https://vop.example.com/mcp \
--header "Authorization: Bearer vopt_..."

Other clients take the same URL and header in their own configuration format. Clients that use OAuth discover the protected resource at /.well-known/oauth-protected-resource.

Tools​

The endpoint is read-only. Every tool respects the user's scope, and none of them starts a scan or changes a setting.

ToolAnswers
search_findingsLists findings, with the filters of the findings page.
vulnerability_summarySummarises the vulnerabilities of the inventory, by severity.
list_packagesLists the packages of the inventory.
list_assetsLists the hosts and containers that report an inventory.
list_agentsLists the agents and their status.

Use a personal token for each assistant, name it after the assistant, and revoke it when the assistant is retired.