quality-validation

Task completion verification compares final AI outputs against initial user prompts to confirm all requirements are met. It enforces that partial completion is not considered done and guides the AI to fix issues or report honestly if blocked.

Ensuring code changes fulfill the user's request involves checking if the code functions as requested, passes applicable tests, and adheres to existing patterns before the AI reports the task as done. This prevents incomplete or incorrect deliverables.

Task completion verification supports documentation and research tasks by providing specific validation steps for each. It compares the final output against the initial prompt to confirm all requirements are met across various multi-format task types.

Stopping partial completion from being considered done requires a validation process that strictly compares deliverables against original requirements. It guides the AI to fix issues or report honestly if blocked, preventing incomplete work from passing.

You can use quality validation for refactoring tasks as it provides specific validation steps for this task type. It verifies that the refactored deliverables precisely match the user's original request, preventing incorrect code changes from passing.

When AI cannot fully meet the original requirements, the validation process guides it to report honestly if blocked. It enforces that partial completion is not considered done and directs the AI to fix the outstanding issues before reporting completion.