tool-selection-framework

Tool selection depends on codebase size, task complexity, and context window needs. Claude Code suits larger codebases and deep reasoning; Gemini CLI handles focused tasks with tighter constraints. The framework evaluates these factors to allocate the optimal tool for each phase.

Multi-phase workflows leverage different tools efficiently by matching each phase to tool strengths. Use exploration phases with one tool, implementation phases with another, based on context limits and reasoning depth required for each stage.

Larger codebases approaching context limits benefit from Claude Code's extended window and reasoning capacity. Smaller, focused tasks fit Gemini CLI's constraints. The framework assesses scale to prevent wasted context and optimize cost versus performance trade-offs.

Exploration tasks require broad reasoning and pattern discovery; implementation tasks need focused execution. Tool selection distinguishes between these types—exploration often benefits from deeper reasoning budgets, while implementation prioritizes efficiency and context precision.

Yes, the framework specifically targets projects with unclear requirements. It provides a structured decision process that evaluates context window size, task complexity, and reasoning depth to identify the best tool allocation before committing resources.

Contextual optimization matches tool capabilities to task scope, preventing unnecessary context consumption. The framework guides phase planning so each tool's window is used efficiently for its assigned work, reducing overhead and improving performance.