new-system

When this skill is invoked:

1. Parse the system name from the user's message. If the user wrote /new-system SpawnSystem, the name is SpawnSystem. If no name was provided, ask for one. The name should be PascalCase and end with System, Manager, or Controller (e.g., SpawnSystem, InventoryManager, InputController).

2. Ask for the following inputs before generating anything:

   • Responsibilities: "What does this system do? Describe in 1-2 sentences."
   • Dependencies: "Does it interact with other systems? If so, which ones?"
   • Data-driven?: "Should it use ScriptableObjects for configuration data? (e.g., spawn rates, inventory slot counts, damage curves)"

   Wait for the user to answer all three before proceeding.

3. Read knowledge base docs for architecture guidance:

   • ${CLAUDE_PLUGIN_ROOT}/knowledge/02-scriptableobjects-data-driven.md — for ScriptableObject patterns
   • ${CLAUDE_PLUGIN_ROOT}/knowledge/03-unity-design-patterns.md — for system design patterns

4. Generate the system class following these conventions:

   • Namespace: [ProjectName].Systems (or appropriate sub-namespace)
   • MonoBehaviour if it needs lifecycle (Update, OnEnable) or scene presence
   • Static class if it's a pure utility with no state
   • [SerializeField] private for inspector fields, never public
   • Events (System.Action or ScriptableObject event channels) for cross-system communication
   • No Find(), FindObjectOfType(), or SendMessage() — use dependency injection or serialized references
   • No method exceeds 40 lines
   • Manual for loops — no System.Linq
   • Include XML doc comments on the class explaining its purpose

5. Generate the ScriptableObject config (if data-driven):

   • Separate [CreateAssetMenu] class for configuration data
   • Name it <SystemName>Config (e.g., SpawnSystemConfig)
   • Include sensible default values
   • Add [Tooltip] attributes on fields

6. Reference assembly definitions: if the system belongs to a specific layer (Runtime, Infrastructure, etc.), mention which .asmdef it should live under, per the 4-layer architecture in KB doc 01.

7. Write the files using the Write tool to the appropriate location in the user's project. Ask the user for the target folder if unclear.

8. Summary: list the generated files, their purpose, and any manual steps the user needs to take (e.g., creating the ScriptableObject asset in the Editor, wiring references).