request-optimizer

Token efficiency starts with analyzing your request before execution. Request-optimizer automatically assesses request specificity, exploration needs, and task decomposition to recommend the right strategy and model choice (Haiku, Sonnet, or Opus), eliminating wasted context and unnecessary processing.

Request-optimizer uses a 5-point analysis framework to automatically decompose complex requests into manageable subtasks, coordinate tool usage, and suggest whether to use exploration agents or specialized tools. It recommends concrete execution steps before you commit resources.

Approval gates in request-optimizer prevent unintended expensive operations by requiring your sign-off on heavy tasks like codebase exploration or model upgrades. You review the recommended strategy first, then approve only what you need.

Yes. Request-optimizer handles vague requests by analyzing what exploration is actually needed, recommending whether to use an Explore Agent, suggesting task phases, and delivering a focused strategy. It turns unclear requests into concrete, efficient execution plans.

Request-optimizer's 5-point analysis includes automatic model selection, recommending Haiku for lightweight tasks, Sonnet for balanced workloads, or Opus for complex reasoning. The recommendation is based on your request's actual requirements, not default assumptions.

Request-optimizer integrates with Explore Agent, TodoWrite, and MCPs to deliver a complete strategy. It analyzes your request, recommends which tools fit best, and prepares coordination steps so all components work together efficiently.