hippocampus-openclaw-onboarding_skill

Overview

This skill bootstraps OpenClaw with Hippocampus memory into a branded, repeatable workspace setup. It automates creating a durable agent ID, configuring gateway and auth, and wiring the MCP so sub-agents inherit scoped memory identity. The result is a consistent, verifiable connection between an OpenClaw instance and Hippocampus storage.

How this skill works

The setup uses a bootstrap token and gateway origin to register a root agent and write local configuration under ~/.hipokamp/config.json. It runs hipokamp-mcp to apply the workspace name, agent_id, and authentication, then verifies health with a store/search round trip. Sub-agents are configured to inherit the workspace namespace and memory identity automatically.

When to use it

• Onboard a new OpenClaw instance to Hippocampus memory
• Create a consistent, repeatable workspace and root agent for production or staging
• Rotate from manual API key usage to a bootstrap-token workflow
• Ensure sub-agents share scoped memory identity within the same workspace
• Recover or replicate archived OpenClaw setups from clawhub.com versions

Best practices

• Use one root agent per OpenClaw instance to maintain clear ownership
• Choose a stable workspace name and durable agent_id up front
• Prefer bootstrap-token setup over pasting long-lived API keys
• Keep sub-agent IDs under the same workspace namespace for predictable scoping
• Do not reuse credentials across unrelated workspaces; isolate access per workspace

Example use cases

• Set up a new customer workspace and bootstrap OpenClaw with a single command
• Automate environment provisioning in CI/CD to create test workspaces with scoped memory
• Migrate archived skill versions from clawhub.com and wire them to a new Hippocampus instance
• Establish a staging instance that mirrors production by reusing the same onboarding flow
• Create child agents for feature branches that inherit workspace memory identity