meta-llm-type

To decide on an AI component type, you diagnose whether the feature relies on knowledge-based Skills, reasoning-based Agents, or access-pattern Commands. This categorization guides your architectural decisions before you start coding the feature.

In LLM architecture, an Agent handles reasoning-based tasks, while a Skill provides knowledge-based capabilities. Distinguishing between them ensures you select the optimal design pattern for structuring AI coding assistants.

You need a Command wrapper for an LLM Agent when platform constraints require specific access patterns. Wrapping the Agent in a Command ensures the component interacts correctly within the broader architectural design.

Yes, you can use component diagnosis for automatically generating documentation from code. It evaluates the feature request and recommends whether the implementation should be structured as a Skill, Agent, or Command.

The best way to structure AI coding assistants is to categorize features into appropriate component types like knowledge-based Skills and reasoning-based Agents. This approach ensures optimal design patterns and clear architectural boundaries.