openclaw

Summary

This workspace provides an AI assistant environment with integrated monitoring tools and expert agent configurations. It is structured to streamline the setup, management, and collaboration of AI agents in operational, design, development, deployment, and audit roles within OpenClaw.

Included Assets

• Core agent configuration files:
  • AGENTS.md, SOUL.md, MEMORY.md, TOOLS.md
• Expert agent workspaces:
  • experts/audit-xiaopi/ (audit/quality expert)
  • experts/deployment-xiaopi/ (deployment/ops expert, with Kubernetes tooling)
  • experts/design-xiaopi/ (design/UX expert)
  • experts/development-xiaopi/ (development expert)
• Kubernetes resources and guidance:
  • experts/deployment-xiaopi/skills/kubernetes-devops/README.md
  • experts/deployment-xiaopi/skills/kubernetes-devops/k8s-manifest-generator/README.md

How to Use

1. Getting Started

   • Review the top-level AGENTS.md for workspace orientation.
   • Each expert agent directory (under experts/) contains its own AGENTS.md, SOUL.md, MEMORY.md, and TOOLS.md files for configuration and operations.

2. Session Workflow

   • Before a working session, read the agent's SOUL.md (identity/goals) and recent memory/ entries for context.
   • In direct sessions (main), also review long-term memory in MEMORY.md.
   • Use TOOLS.md for environment- or device-specific configurations.
   • Follow role-specific principles and communication guidelines as described in each expert’s SOUL.md.

3. Expert Roles

   • Audit (审核小皮): Oversees quality/standards for deliverables.
   • Deployment (部署小皮): Handles system deployment, ops, and monitoring, including Kubernetes use.
   • Design (设计小皮): Responsible for UX, visual, and interaction design.
   • Development (研发小皮): Developer area (see files for specifics).

4. Kubernetes Skills

   • Use the deployment expert’s skills directories for:
     • Generating manifests
     • Setting up workloads and services
     • Following cloud-native production best practices

Notes

• Each expert agent manages its own memory, tools, and identity files. Contextual information is siloed to avoid leakage between roles or group sessions.
• Files may contain both English and Chinese documentation, reflecting the team’s practices.
• No executable code is present; the workspace is configuration- and documentation-oriented.
• Long-term and session memory management is emphasized—important decisions or updates should always be saved to the relevant MEMORY.md or memory/YYYY-MM-DD.md files.
• Role boundaries and collaboration workflows are clearly defined to avoid overlap or confusion among expert agents.

For any role-specific usage, always consult the corresponding expert workspace.