Subagent-Driven Development FAQs

Question 1

How does the two-stage review work?

Accepted Answer

Every task is reviewed twice: first, a reviewer subagent checks if the implementation matches the specification exactly. Once passed, a second review assesses code quality, maintainability, and test coverage.

Question 2

What is Subagent-Driven Development?

Accepted Answer

It is a specialized workflow where a main AI agent dispatches temporary 'subagents' to handle specific tasks within a larger plan, ensuring clean context and focused execution for each component.

Question 3

Does it support Test-Driven Development (TDD)?

Accepted Answer

Yes, subagents dispatched via this skill are instructed to follow TDD patterns, writing and passing tests for every task before proceeding to the review stage.

Question 4

Why use this instead of standard plan execution?

Accepted Answer

This skill keeps you in your current session while providing 'fresh' subagents for every task, preventing the AI from getting confused by previous steps or irrelevant context.

Question 5

What happens if a reviewer finds an issue?

Accepted Answer

The implementer subagent is tasked with fixing the issue immediately, and the reviewer re-evaluates the code. This loop repeats until the code passes both compliance and quality gates.

Subagent-Driven Development

主要功能

使用场景

Subagent-Driven Development

主要功能

使用场景