Agents

Overview#

The Agents page is your central registry for all OpenClaw agents. Agents are auto-discovered when you sync a connector. Each agent shows its health status, runtime configuration, model settings, skills, and workspace files.

You can enable/disable agents, trigger the kill switch for emergency stops, inspect the core configuration tabs plus evaluation, scorecard, and governance views, and edit workspace files with conflict detection.

Prerequisites#

OpenClaw Configuration#

Agents are defined in your OpenClaw configuration. Each agent has an identity, model configuration, skills, tools, channels, and cron jobs.

Key configuration files in OpenClaw:

• -Agent identity: name, description, emoji, and role defined in the OpenClaw config YAML
• -Model config: primary model and optional fallbacks (e.g., anthropic/claude-sonnet-4-6)
• -Skills: agent capabilities that can be enabled/disabled
• -Workspace files: custom instructions and data files in the agent's working directory

ClawButler Configuration#

After syncing from a connector, agents appear automatically. You can:

• -Enable/Disable: Toggle agent availability. Disabled agents stop receiving tasks.
• -Kill Switch: Emergency stop that immediately disables the agent locally AND propagates to OpenClaw.
• -Delete: Remove an agent from ClawButler (does not affect OpenClaw).

Expected Results#

After syncing, the Agents page shows:

• -Grid or list view of all agents with health status dots (green=healthy, amber=attention, red=error)
• -Agent avatar with identity emoji and name
• -Runtime status indicators (online, disabled, killed)
• -Connected instance name and agent count badge

Click any agent to inspect the configuration tabs plus evaluation, scorecard, and governance views.

Web Operations#

The Agents page supports two view modes:

Grid View (Default)

Cards showing avatar, name, health status, and description. Best for scanning many agents at a glance.

List View

Compact rows with avatar, name, connector, and health dot. Better for dense lists. Toggle with the view mode button.

Filtering

Use the health filter pills (All / Healthy / Attention / Error) to narrow the list. View mode is persisted in localStorage.

Agent Detail

Click an agent card to inspect the configuration tabs and the governance view. The Governance tab summarizes recent decisions, activity, and traces, and exposes pause / resume / retire lifecycle actions. The Files tab lets you view and edit workspace files with conflict detection.

CLI Operations#

List all agents with optional health filter

$ ap agent list

Show details for a specific agent

$ ap agent show <id>

Review recent decision, activity, and trace summaries for one agent

$ ap agent governance <id>

Pause, resume, or retire an agent from the lifecycle controller

$ ap agent pause <id> --reason "Maintenance window"
ap agent resume <id> --reason "Window closed"
ap agent retire <id> --reason "Decommissioned"

Enable a disabled agent

$ ap agent enable <id>

Disable an agent

$ ap agent disable <id>

View agent configuration (overview, files, tools, skills, channels, cron)

$ ap agent config show <id>

View or update model configuration

$ ap agent config models show <id>

Mobile Operations#

The Agents tab is the second tab in the bottom navigation. Features include:

• -Search bar for filtering by name or platform
• -Health filter pills (All / Healthy / Attention / Error)
• -Tap any agent to view details on a dedicated screen
• -Pull-to-refresh to reload agent data

FAQ#