AI Agent

Following our previous discussion on the RoleEngine Core, it’s time to bridge the gap between a raw CLI “brain” and a real-world communication platform: WeChat. In VisAgent, we don’t believe in heavy, bloated frameworks. Instead, we use a CLI-Native Bridge pattern. This post explores how we connected the “Clawbot” (our WeChat interface) to a hand-rolled Gemini CLI kernel. The Request Flow: From Chat to CLI The architecture is a chain of specialized tools, each doing one thing well. Here’s how a message travels from your phone to the AI: ...

[!TIP] Objective: Implement “zero-intrusion” architectural document synchronization and an AI-driven system self-repair closed-loop. 1. Zero-Intrusion Metabolism (Metabolic Governance) In modern Agent development, maintaining an ARCHITECTURE.md and routing table is often a developer’s nightmare. Our solution is a Metadata-Free “metabolism” mechanism: Perceive Changes: Mounted on a Git hook, it captures code variations via Git Diffs. Semantic Inference: The AI automatically analyzes code intent to update the machine-readable .anti_bot_map.md (RAG routing table) and the human-readable ARCHITECTURE.md (Architectural Lore). Strong Consistency: Completely eliminates the cost of manual documentation maintenance; docs are code, and code is the map. 2. The System’s Pulse: Heartbeat Engine To ensure Agent stability in unattended environments, we configured an independent Heartbeat process. It’s more than just resource monitoring; it’s a “self-awareness” probe: ...

[!TIP] Objective: Enable multi-modal (vision/files) perception via CLI parameters and build a dynamically-loaded plugin-based skill system. 1. Letting the CLI “See” the World Multi-modal capabilities don’t necessarily require complex SDKs. In VISAGENT, we leverage the native support for file paths in the gemini CLI (@path) to achieve sensory integration at the RoleEngine layer: def _do_raw_invoke(self, message, files=None): # Construct multi-modal suffix mm_suffix = "" if files: mm_suffix = "\n" + "\n".join([f"@{f}" for f in files]) # Append to final Prompt full_input = f"{message}{mm_suffix}" # ... execute subprocess Field Experience: To handle complex visual tasks, we encapsulated a dedicated vision_expert skill. By using the DEEP reasoning mode, we guide the AI through Chain-of-Thought thinking, enabling precise identification of screenshots and UI components. ...

[!TIP] Objective: Evolve from single-step prompts to a “Plan-Simulate-Execute” closed-loop, introducing YAML-based DAG asynchronous orchestration. 1. From “Chatting” to “Execution”: The AX (Architect-Executive) Paradigm The watershed between a simple Chatbot and an Agent is the ability to decompose vague goals into actionable steps. In VISAGENT, we implemented the AX Planner logic: Architect: Receives requirements and outputs a TODO.json. Execution is forbidden; only planning is allowed. Simulation: Before execution, a “Security Expert” role performs a risk assessment on the plan. Executive: Executes each step via execute_step. This “think before you act” mechanism is easily implemented on any CLI using simple Prompt constraints: ...

[!TIP] Objective: Build an Agent engine with state management and isolation by wrapping a CLI (e.g., gemini CLI) without framework reliance. 1. Why CLI Over Frameworks? VISAGENT’s philosophy is CLI-Native. Directly wrapping a CLI (like gemini CLI) offers superior control and transparency: Atomicity: Each call is a single inference step. State Transparency: Session resume/suspension is strictly controlled by paths and the Resume flag. Isolation: Permission sandboxing is achieved through native OS tools. 2. Core Wrapper: RoleEngineBase Execution is triggered via subprocess. To prevent Shell parameter overflow from long Prompts, we use stdin for transmission: ...