Mentions and commands
Look up what @ and / open in the message box, what each item does, and the keys that drive both menus.
Reference Updated 20 Nov 2026What this covers
Two characters open a menu in the message box. The @ character names something for Panda to use: an agent, an app, a skill, an integration or a file. The / character runs a command or a skill. This page lists what each menu holds and what picking an item does.
The @ menu
Type @ at the start of the message or after a space. The menu is headed Mention and shows up to 15 matches, in these groups:
| Group | What it lists | What picking it does |
|---|---|---|
| Agents | The open project's agents that you call by name | Panda takes on that agent's instructions for this message |
| Apps | Your apps, connected or not | Panda uses that app for this request. If the app is not connected, Panda offers to connect it |
| Skills | Skills you call yourself, from this project and from all projects | Panda follows that skill for this request |
| Connectors | The MCP servers switched on for this project or for every project, except those an app already covers | Panda uses that connector's tools for this request |
| Files | Every file in the project, not folders | The file's path goes with your message, so Panda knows exactly which file you mean |
An agent, app, skill or connector becomes a chip on top of the message box. There is one at a time, so picking another replaces it, and the chip's x removes it. Files become attachment chips. Add as many as you need. Your message then shows as @name followed by what you wrote.
The groups always show in the order of the table. Inside each group the best matches come first. An exact name ranks first, then one that starts with what you typed. After that comes one that contains it, then a match anywhere in the path.
The / menu
Type / as the first character of the message. The menu is headed Commands and has two groups, Skills first, then Built-in.
| Built-in | What it does |
|---|---|
| /model | Switch model mid-session. Opens the model list. |
| /plan | Enter plan mode. The chat plans before it changes anything, see Plan mode. |
Skills lists the skills you call yourself, from this project and from all projects, and the commands the chat's engine offers. An engine's commands appear once the chat has started, so in a new chat they show after your first message. With Codex and OpenCode that includes /compact, which summarizes the conversation to free up room.
Picking a skill or an engine command sends it straight away, as a message of its own.
Keys
| Key | In the @ menu | In the / menu |
|---|---|---|
| Up and Down | Move through the list, wrapping round at the ends | Move through the list, wrapping round at the ends |
| Enter | Pick the highlighted item | Pick the highlighted item |
| Tab | Pick the highlighted item | Nothing |
| Escape | Close the menu | Close the menu until the message no longer starts with / |