> ## Documentation Index
> Fetch the complete documentation index at: https://docs.workser.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# The AI agent panel: everything one AI agent holds

> Open one AI agent's panel in Chat: Profile, Today, Autonomy, Resources, Automations and Files, plus the 12-section Manage dialog and what appears only when the agent has it.

**In a one-to-one chat with an AI agent, its name at the top of the chat opens a panel on the right with everything that AI
agent holds.** The panel starts closed. Use the name pill (tooltip "About \<name>") or the bot button beside it
("About \<name>: computer, resources, profile, automations and files").

There are two doors to the same AI agent:

* **The panel** opens from a chat and is arranged for working with that AI agent: its screen now, its limits, its files.
* **The Manage dialog** opens from the gear on the AI workforce page or a card, and from **Settings › AI workforce**. It is
  arranged for setting the AI agent up, in 12 sections.

## The six tabs

| Tab | What it shows |
| - | - |
| **Profile** | How it works: instructions preview, its model, skills, connectors, and a Memory row ("n facts · m to check"). |
| **Today** | Its screen now, then **Now** (its state), **Recent** Tasks, **This week** (Tasks, credits, stuck) and **Results**. |
| **Autonomy** | Probation and what it may do without asking. Shows only for an AI agent hired with a staff record. |
| **Resources** | Its computer, browser and logins, phone, email, card, connectors, tools, chat apps and software. |
| **Automations** | Work it owns: on a schedule or when something happens, plus what it suggested. |
| **Files** | Software it built and the Library, with **Open Library**. |

**Why a tab may be missing.** Phone, Email, Cards, Browser, Setup and Connectors appear only when the AI agent actually holds
that thing. Nothing is lost: **Resources** always lists every resource with its state ("Ready", "Not set up"), and what the
AI agent lacks is offered on **Profile**.

**Two shortcuts that go to Manage.** Tapping **Email** or **Phone** in the panel now opens the Manage dialog on **Identity**,
because a work email and phone number are managed in one place. The panel's own Email and Phone views are no longer entered
from here.

## The Manage dialog's 12 sections

| Section | What it controls |
| - | - |
| **Overview** | How it is doing, what it costs and what it did lately; **Give a task**. |
| **Expert kit** | Everything that makes it good at its job here, what its role would add, and every change with Undo. |
| **Profile** | Its name, label, description and how it looks. |
| **Instructions & model** | How it works, the model it thinks with and what it can do. |
| **Identity** | Its own work email, phone, saved logins and card. |
| **Computers & browser** | Where it works: its computer, workspace and browser. |
| **Connectors & skills** | The connectors, skills and plugins it may use. |
| **Knowledge & files** | Price lists, policies, examples of good work: files, notes and links it reads. |
| **Automations** | Work it owns and does on its own. |
| **What it keeps** | Its memories, the customers it knows, its skills, and its computer and workspace. |
| **Rules & alerts** | When it asks you first, and how it tells you about its work. |
| **Advanced** | Limits, role presets, keys and passwords, helpers, and removing it. |

Two sections are drafts saved together: **Profile** and **Instructions & model**. Everything else saves as you change it.
Until you press **Save changes**, the footer says "Unsaved changes. New tasks use them once saved; running tasks keep their
setup." After a save: **"Saved. New tasks use it."**

If a save lands but has not reached new tasks yet, Workser says so: "Saved. It isn't applied to new tasks yet; your next
saved change applies it."

## An AI agent's state, in plain words

**Today** and the sidebar use the same words everywhere: **Request** (it needs you), **Working**, **Waiting to start**,
**Paused**, **Done**, **Stuck** and **Stopped**. "Ready" means free, never "done". Hovering an AI agent's name in the sidebar
shows its last Task and up to four recent steps; a working one opens a small live card after a moment, never a second live
connection.

## Removing an AI agent

**Advanced › Remove** is for an owner or admin; a member never sees it. The confirmation says what goes and what stays:
it leaves the team and its Automations stop, its work email and phone number are deleted, and its past Tasks and emails
stay in your history.

## Related

* [Skills and agent kits](https://docs.workser.ai/using/skills-and-kits.md)
* [Memory](https://docs.workser.ai/using/memory.md)
* [AI models](https://docs.workser.ai/using/models.md)
* [Autonomy and probation](https://docs.workser.ai/using/autonomy-and-probation.md)
* [Sharing and handoffs](https://docs.workser.ai/using/sharing-and-handoffs.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.