clawd-presence_skill

Overview

This skill provides a compact terminal-based presence display for AI agents, showing a customizable monogram (A-Z), a color-coded status state, and the agent's current activity. It offers faster, glanceable feedback than chat by broadcasting agent state continuously on a dedicated screen. The display is designed for always-on visibility so teams can monitor agent activity without opening a UI or reading logs.

How this skill works

The agent updates its state by calling a lightweight status script with a state and optional message; the display script runs in a dedicated terminal and renders a monogram, the current state, and activity text. States are color-coded (idle, work, think, alert, sleep) and the system auto-idles after a configurable timeout to avoid stale displays. Configuration can be auto-detected or set manually, including monogram letter, agent name, and timeout.

When to use it

• When you need always-on, glanceable visibility of what an agent is doing.
• On a operations dashboard or a developer workstation to reduce chat polling.
• During demos or monitoring sessions where fast feedback beats detailed logs.
• In multi-agent setups to quickly distinguish individual agent activity.
• When you want a low-overhead presence display for remote or local agents.

Best practices

• Assign each agent a unique monogram letter (A-Z) and optional short name for quick identification.
• Use concise activity messages so the display remains readable at a glance.
• Set an appropriate auto-idle timeout to prevent stale states (or disable if not needed).
• Color-code states consistently across your team to avoid confusion.
• Run the display in a dedicated terminal or screen that remains visible to observers.

Example use cases

• Run a display on a wall-mounted monitor to show active agents across a workspace.
• Place a small terminal window for each agent on an operations laptop for rapid status checks.
• Integrate status updates into agent workflows so tasks broadcast when they start and finish.
• Use alert state to signal when human attention or intervention is required.
• Archive visual presence logs alongside agent artifacts for auditing and debugging.