What problem does it solve? It explains how to configure the agentmemory server, including which environment variables exist, which features are off by default, and how the port layout works, so you avoid misconfiguration and unexpected token costs. ## Core Features & Use Cases - Environment Variable Reference: Lists all 42 recognized AGENTMEMORY_* variables read from the environment or ~/.agentmemory/.env. - Default Behavior Guidance: Documents that no API key is required, that zero-LLM mode uses BM25 plus local embeddings, and that token-spending features like AUTO_COMPRESS and INJECT_CONTEXT ship off. - Port Layout Explanation: Describes the port quartet anchored at REST 3111 with streams, viewer, and engine offsets, relocatable via --port or --instance. - Use Case: When enabling auto context injection or setting an Anthropic API key, consult this Skill to write the correct KEY=value lines and restart the server. ## Quick Start Ask the AI to show which agentmemory environment variables to set in ~/.agentmemory/.env to enable auto-compress with an Anthropic API key.