executing-plans

This skill enables the AI agent to execute detailed implementation plans provided in a separate session. It emphasizes a batch execution approach with frequent checkpoints for review, ensuring adherence to the plan and providing opportunities for adjustments based on architect feedback. The core principle is to minimize errors and maintain alignment throughout the development process.

The skill involves loading the plan, critically reviewing it, executing tasks in batches (defaulting to the first 3), reporting progress with verification outputs, and continuing execution based on feedback. It also includes clear guidelines on when to stop and ask for help, preventing the agent from proceeding with unclear instructions or encountering blockers.

When all tasks are completed and verified, the skill leverages the `finishing-a-development-branch` sub-skill to finalize the development, ensuring all tests pass and presenting options for final adjustments.

What It Does

Loads a written implementation plan, executes tasks in batches with verifications, reports progress for review, and continues based on feedback.

When To Use

When you have a detailed, step-by-step implementation plan to execute, and you want to ensure accuracy and alignment with architect's vision through regular checkpoints.

Installation

Copy SKILL.md to your skills directory