MCP servers ​

The MCP Servers view connects external tool servers over the Model Context Protocol — scanners, trackers, your team's services, or custom scripts — so the agent can call them as tools without everything being pasted into the prompt.

The MCP Servers view: the built-in public-intel server connected, and a form to add a local (stdio) or remote (HTTP) server

  • A built-in public-intel server ships enabled, giving the agent security-research tools out of the box.
  • Add your own servers, either local (stdio) — a command the app spawns — or remote (HTTP).
  • Connected tools are namespaced mcp__<server>__<tool> and merged into the agent's tool list. Live status and each server's tool list show in the view.

Approvals ​

Each MCP tool's first use asks for approval, showing the call. Choose Always allow to trust a tool so it won't ask again. This reuses the same approval gate as the rest of the app — see Approvals & safety.

See also ​

Hacker Sidekick documentation