writing-plans

To turn feature requirements into executable implementation steps, you can generate a plan that translates requirements into concrete, bite-sized development tasks with file-level scope and TDD verification. This maps exactly what to read, change, test, and commit.

The best way to plan multi-step feature delivery for an unfamiliar codebase is to use execution-ready documentation that assumes little context. This approach explicitly maps exact file paths, commands, expected outcomes, and verification steps so execution is unambiguous.

To structure task decomposition using a TDD workflow, break work into small 2–5 minute steps following a failing-test, minimal implementation, and passing-test loop. This ensures each task has testing and verification baked in.

Yes, you can use implementation plans to coordinate subagent-driven development. The plans provide execution handoff workflows with exact paths and expected outcomes, making it easy to delegate bite-sized tasks to subagents even in context-sparse scenarios.

Yes, this approach satisfies requirements for DRY and YAGNI guidance during refactoring tasks. It applies to multi-step feature delivery and refactoring by generating concrete, file-specific steps with exact run and commit commands.

You should avoid using bite-sized implementation plans for simple, single-file edits that require no task decomposition or testing loops. The overhead of mapping exact paths, commands, and verification steps is best suited for multi-step feature delivery and context-sparse scenarios.