Skills, memory and agents

One settings screen shapes how the assistant behaves everywhere: what it remembers, what it can reach, and what it is allowed to do without asking first.

Skills. Custom, reusable prompts you invoke by name, or that fire automatically when a workflow calls for them.
Memory. Durable facts the assistant keeps across conversations, all visible, all removable, nothing hidden.
Automations. Set up event-based rules that trigger actions automatically when something happens in your workspace.
Agents. Purpose-built AI personas, each with its own system prompt, its own model choice, and a narrowed set of tool permissions.
Connected tools. Link external developer tools and databases (over the Model Context Protocol) to extend what the assistant can reach.
Permissions. Decide per tool whether it waits for your explicit approval or just gets on with it.