implement-plan

To execute a phased implementation plan, you implement each phase fully while updating the plan's checkboxes to reflect completed work. This approach translates an approved technical roadmap into working code with verifiable progress throughout the software development lifecycle.

When codebase reality diverges from the implementation plan, phase-based execution stops immediately. It generates a structured issue report describing the mismatch and asks how to proceed, preventing incorrect code changes from cascading through subsequent phases.

To validate code changes against success criteria without ad-hoc bash tests, use a dedicated test-runner agent. This ensures automated pytest and linting checks run consistently, catching failures before proceeding to the next implementation phase.

Yes, an approved technical plan file is required before starting phase-based implementation. The process requires reading the plan and its referenced files fully, then executing unchecked phases while following intent and adapting to the actual codebase environment.

The best way to manage multi-phase feature implementation is executing an approved plan phase-by-phase while validating each stage against success criteria. This method maintains alignment with the original roadmap while adapting to codebase reality and verifying results.