GSD Plan-First Orchestrator FAQs

Question 1

Does it perform the research itself or use subagents?

Accepted Answer

It acts as a lean orchestrator, delegating heavy research and analysis to subagents to keep your primary conversation context clean and focused on the plan.

Question 2

What is the GSD project requirement for this skill?

Accepted Answer

This skill requires a standardized .planning/ directory, typically initialized via the /fh:new-project command, to track roadmaps, state, and architectural decisions.

Question 3

How does it use historical context?

Accepted Answer

If claude-mem is available, it searches past sessions for relevant 'gotchas', decisions, and trade-offs to ensure past mistakes aren't repeated in the new plan.

Question 4

How does it handle complex technical tasks?

Accepted Answer

It evaluates tasks against a complexity matrix and suggests either inline research or dispatching a deep-research subagent to find patterns and pitfalls before you start coding.

Question 5

Why does it suggest splitting plans?

Accepted Answer

Splitting large phases into focused 3-4 task plans ensures that Claude operates at peak quality within its context window, reducing errors and rework.

GSD Plan-First Orchestrator

主要功能

使用场景

GSD Plan-First Orchestrator

主要功能

使用场景