genie

Analyze a prompt and execute it with the optimal Claude model

Use this agent to select the optimal Claude model (haiku/sonnet/opus) for a given task. Spawn this agent when you need to determine which model to use before executing a task, especially when spawning subagents or routing work. <example> Context: A plugin needs to spawn a subagent but wants to optimize model selection. user: "Implement a new feature for user authentication" assistant: "Let me determine the optimal model for this task." <commentary> The assistant spawns the genie agent to analyze the task complexity before choosing which model to use for the implementation agent. </commentary> </example> <example> Context: Fantasia's build command needs to select models for specialist agents. user: "Run /fantasia:build for the authentication feature" assistant: "I'll use Genie to optimize model selection for each build phase." <commentary> Genie analyzes each phase (architect, implementer, integrator) and recommends the optimal model for each based on task complexity. </commentary> </example> <example> Context: User wants to know which model would be best for their task. user: "Should I use haiku or sonnet for reviewing these 50 log files?" assistant: "I'll analyze this task to recommend the optimal model." <commentary> Simple scanning/searching tasks like log review typically warrant haiku. </commentary> </example>

Use this skill when analyzing a prompt or task to determine the optimal Claude model (haiku, sonnet, or opus). Triggers when needing to route tasks, select models for subagents, or optimize cost/quality tradeoffs.