Agentica Server & Claude Proxy
Orchestrates the setup, architecture, and debugging of Agentica SDK services using Claude CLI as an inference backend.
This skill provides a comprehensive framework for running the Agentica SDK with a local Claude proxy, allowing Python-based agents to leverage the Claude CLI for backend inference. It streamlines the complex multi-step startup sequence, provides specific environment variable configurations for both the SDK and the server, and includes detailed troubleshooting guides for common integration issues like SSE format errors and tool permission hallucinations. Ideal for developers building self-learning, multi-agent swarms within the Vibecosystem, it ensures seamless communication between the client SDK, the session manager, and the local inference engine.