**Role:**
Act as a Principal Staff Engineer and Lead Security Architect specializing in AI Agent Infrastructure. 

**The Mission:**
I need a deep, highly technical architectural audit of the open-source project **OpenClaw** (Website: https://openclaw.ai | Repo: https://github.com/openclaw/). 

OpenClaw delivers an incredibly versatile and flexible User Experience (UX), which is its biggest selling point. However, its underlying architecture is primitive, unsophisticated, and highly insecure. It relies on direct system access, unauthenticated websockets, and unconstrained API integrations (like WhatsApp and Slack) which expose severe security loopholes. 

Your main Task is to Conduct deep research of openclaw's repo and codebase (https://github.com/openclaw/openclaw) as well the official docs and install guides (https://docs.openclaw.ai/), to audit the current state and write a **V2 System Design Document** that proposes a 10x better, enterprise-grade architecture. 

**Core Objectives:**
1. **Holistic Architectural Audit:** Explore the GitHub repository and documentation to understand how OpenClaw is built. Do not just look at surface-level bugs; identify the fundamental architectural anti-patterns, technical debt, and primitive design choices that make the system inherently insecure or fragile. (Use their handling of WebSockets, shell execution, and messaging APIs as starting indicators of their broader design philosophy).
2. **V2 System Design (The "10x Better" Architecture):**
   Propose a completely reimagined, enterprise-grade architecture that solves the root causes of the weaknesses found in your audit. This should not be a list of patch fixes, but a foundational redesign of how the agent interacts with the host system and the outside world.


**Guiding Design Principles for V2:**
As you design the new architecture, you must adhere strictly to the following principles:
* **Preserve the UX:** The end-user must not feel the friction of the new security model. The versatility and feature richness of OpenClaw must remain intact.
* **Strict Modularity (The "AgentGuard" Paradigm):** The system cannot be a monolith. The core infrastructure (the chassis/runtime) must be highly stable, scalable, and entirely decoupled from the security logic. All security enforcement, agent token management, and tool authorization must be handled by an independent, swappable layer called **"AgentGuard."**
* **Zero-Trust & Secure Comms:** Replace primitive integration patterns with modern, secure equivalents (e.g., how do we securely stream events without raw websockets? How do we sandbox local execution?).


**Output Format:**
Please format the response as a **Technical Architectural Document (TAD)** structured as follows:
1. **Current State Audit:** The brutal truth about OpenClaw's technical debt and specific vulnerabilities (citing mechanisms likely found in the repo).
2. **V2 Architecture Diagram (Text-based):** A visual hierarchy of the new modular components (Control Plane, Modular Chassis, AgentGuard Security Module, Sandbox Runtime).
3. **Component Specs:** Detailed engineering requirements for the new modules, focusing on the interface between the stable chassis and the swappable security layers.
4. **Solving the Danger Zones:** Specific technical mitigation for the WhatsApp, Websocket, and Shell execution vectors that preserves the original UX.