pelizzai-execution-plans

To execute an approved technical plan task-by-task, this skill breaks it into verifiable tasks with per-artifact validation and proportional code review, ensuring high-integrity implementation without skipping quality checks.

Human-in-the-loop execution workflow for software engineering agents enforces explicit human ratification of execution strategies, isolation modes, and commit strategies before critical decisions are applied to the codebase.

You cannot use this structured execution workflow without an approved PRD or project plan, as it requires a defined strategy to break down into verifiable tasks for state management and integration.

The best way to manage state and commit strategies during multi-task agent execution is enforcing defined isolation modes and mandatory review lenses to ensure delivery quality and proportional review.

Agents make unauthorized architectural decisions during code implementation when lacking strict execution discipline, which this skill prevents by enforcing human-in-the-loop gates for critical ratification.

You should not use a structured task-based workflow for agentic code review if your project lacks an approved plan or requires rapid prototyping without explicit human ratification of execution strategies.