Workspace Guidelines for AgentRQ
A workspace is where one agent works for you: its tasks, its memory, its skills and its connection. These guides explain how to read a workspace without opening it, starting with the colored dots that AgentRQ draws beside every workspace and every task, and walk through every tab of a workspace's settings, with annotated screenshots.
Workspace Settings, Tab by Tab
Open a workspace and click the gear button in its top right corner. Each tab has its own guide:
- → General: the name, the agent's working directory, its mission and its self-learning strategy.
- → Setup: start the agent on one of your machines, or connect Claude Code, an ACP agent, Antigravity, Codex or DeepSeek Harness yourself.
- → Automations: the tools the agent runs without asking, YOLO mode, and clearing its context before each task.
- → Notifications: which task events notify you, by email, Slack, push or desktop alert.
- → Memories: what the agents wrote down for the next one, and your teammates' memory.
- → Skills: import skills from GitHub, turn them on here, and give them to your team.
- → Sharing: which teams see the workspace, and whether they read its tasks and memory.
- → Slack: a Slack channel for the workspace's updates and approvals.
- → Input: voice input language, automatic task titles and a message send delay.
- → Local Storage: a copy of the tasks on this device for instant lists and offline search.
- → Danger Zone: archive, purge, or reset the workspace ID after a leak.
One Color Scheme Everywhere
AgentRQ uses the same colors in the sidebar, the task lists, the Kanban board and each task's page, so you only have to learn them once. Yellow means it is waiting on you, green means the agent is working or done, red means something is stuck, grey means nothing has started, and cyan means it is scheduled for later. A dot that moves means something is happening right now. Workspace colors and blinking dots explained covers each one.