Summary of Claude Code Ecosystem | Bondar Academy

Lesson Summary

Cloud Code Overview In this lesson, we recap the key components and functionalities of Cloud Code , which is powered by three models: HYCU , SONNET , and OPUS . Models and Intelligence Levels HYCU : Basic model SONNET : Intermediate model OPUS : Most advanced model Each model has varying levels of effort required, influencing how deeply Cloud Code explores tools for context collection. Higher intelligence and effort levels lead to faster token consumption. Configuration Levels User-level configuration : Settings, rules, and skills shared across all projects. Project-level configuration : Specific settings, rules, and skills applied per project. Stateless Nature Cloud Code operates as a stateless system , meaning it does not retain conversation history. Each request sends the entire conversation back to the model for reprocessing. An internal caching mechanism exists for five minutes to optimize token usage. Best Practices Keep the context window small and relevant. Avoid switching tasks within the same session. Set specific goals for each working session. Memory System Cloud Code features a multiple layer memory system with: Cloud.MD : Automatically loaded knowledge base per project or user. Rules : Automatically loaded but can be scoped by project paths. Skills : Knowledge bases for specific tasks, invoked dynamically or manually. This concludes the recap of Cloud Code functionalities and best practices.