startupO-lite

To reduce token usage when starting an agent orchestrator, load minimal context upfront and query RAG on-demand for task-specific knowledge. This approach fetches detailed information only when required, optimizing resource consumption during complex scheduling tasks.

RAG-first context management is an orchestrator startup approach that loads minimal initial context and retrieves task-specific knowledge on-demand. It performs checks on Git status, stack health, MCP/RAG status, and container staleness before reporting active state and available agents.

Initiate orchestrator operations with on-demand RAG queries by using a lightweight startup skill that loads minimal context. It checks Git status, stack health, MCP/RAG status, and container staleness, then reports its active state and available agents for task execution.

Yes, the orchestrator checks MCP and container status during startup. It performs health checks on Git status, stack health, MCP/RAG status, and container staleness before reporting its active state and available agents to ensure a ready environment.

The best way to start a complex scheduling task with minimal context is using a RAG-first orchestrator startup. It ensures the orchestrator has just enough context to understand the request and intelligently fetches detailed rules or priorities from the RAG system as needed.

You should not use a full-context startup when token efficiency is critical. Instead, use a RAG-first approach that loads minimal context upfront and queries RAG on-demand, avoiding heavy initial token consumption while still supporting complex scheduling operations.