writing-plans

Implementation plans are broken down into bite-sized 2–5 minute tasks with explicit file paths, complete code blocks, and exact verification commands. This prevents stalled implementations by ensuring engineers can proceed with minimal guesswork and without searching for missing context.

Test-driven development task framing structures each implementation step as a loop: write a failing test, verify the failure, write the minimal implementation, verify the pass, and commit. This TDD-oriented task framing reduces regression risk during feature development and multi-step refactors.

Yes, implementation planning is specifically designed for delegated work via subagents and situations where the implementer lacks codebase context. It provides file-level specificity and exact commands so subagents can execute tasks without needing domain context or manual codebase navigation.

The best way to plan a multi-step refactor is to generate bite-sized, sequential tasks that include exact file paths, copy-paste-ready verification commands, and DRY/YAGNI guidance. This makes the implementation unambiguous and suitable for frequent commits throughout the refactor cycle.

No, you do not need prior codebase context. The plan provides file-level specificity, complete code blocks, and exact commands so implementers can proceed without searching for missing context or needing deep domain knowledge to navigate the codebase.