Configure OpenClaw Agent

Configure OpenClaw agents from the real runtime layout, not from assumptions. The executor should always know which machine owns the runtime, which workspace belongs to the target agent, and which file controls each behavior.

Trigger boundary

Use this skill when the task involves:

• Setting up a new OpenClaw agent workspace
• Writing or revising workspace files such as AGENTS.md, SOUL.md, IDENTITY.md, USER.md, TOOLS.md, or memory files
• Choosing a tool profile (full, coding, messaging, minimal) for an agent
• Configuring tool allow/deny lists, provider-specific tool restrictions, or per-agent tool policy
• Setting up skill loading, extra skill directories, skill env vars, or bundled-skill allowlists
• Enabling memory tools or deciding where durable behavior should live in the workspace
• Hardening exec approvals, sandboxing, elevated access, or tool-loop detection
• Debugging why a tool is blocked, a skill is not loading, or the wrong workspace is active

Do NOT use this skill for: