Your Assistant's Page
Each assistant you connect has one page in Scout, where you see and change everything it can do. Open it from Personal assistants in the sidebar, or from its card on Home. Its address is the assistant's name: scout.withwillow.ai/agents/instinct, /agents/muse, /agents/grok-bot or /agents/dots.
The header
- Live or Paused. Click Pause to stop the assistant from using any tool, and Go live to turn it back on. A paused agent still counts as your one connection for that assistant.
- Safety and a score. Click it to see what to tighten. See Safety score.
Access
The card at the top says whether the assistant has access: Instinct has access, Grok Bot's access expired, or Muse has no access, for example.
| Fact | What it shows |
|---|---|
| Last connected (or Last signed in for Instinct) | When the assistant last connected. |
| Session | Unlocked until a time, or Locked, with how long one lasts once you approve it. |
| Access ends | When access expires, and how many days are left, or Never. |
| Can use | How many tools and apps it has. |
Click Connect, or Reconnect once it connected before, for the steps to connect the assistant again. Use it when the assistant lost its link or password, or after access expired or you stopped it.
Connectors
The apps and MCP servers the assistant gets, and how each tool runs. Click Add to add one. See Apps and tools.
Settings
How the assistant gets in, how long it stays in, and how it asks you:
- Approvals: where requests reach you. Click Edit for How you approve.
- Password (Instinct) or Sign-in (Muse, Grok Bot and Dots).
- Access: how long access and a session last, the link or server URL, and stopping access.
- Delete agent.
See Access and settings.
Activity
Beside the page, Activity shows what the assistant does, as it happens: every sign-in, session and tool call, newest first. Requests that wait for you show up right in the line that asked, so you can answer them there. See Activity.
Safety
Click the Safety pill in the header. The safety page rates what the assistant can do through this agent: the accounts it uses, what runs without your approval, and how long access lasts. See Safety score and Guards.