gsd-execute-phase

Execute project phase plans in parallel waves with isolated subagents.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/dobrinz123/claude-skills-share --skill gsd-execute-phase-dobrinz123
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-execute-phase
Source: https://github.com/dobrinz123/claude-skills-share/tree/main/skills/gsd-execute-phase
Command: npx skills add https://github.com/dobrinz123/claude-skills-share --skill gsd-execute-phase-dobrinz123

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates the execution of all plans within a project phase, removing manual coordination overhead when many dependent tasks must run in parallel or in staged waves.

Core Features & Use Cases

  • Wave-based Orchestration: Discover plans, analyze dependencies, group tasks into waves, and execute them in parallel or staged sequences.
  • Subagent Isolation: Spawn subagents that load fresh execution context for deterministic plan handling and result collection.
  • Flexible Filters & Modes: Support for --wave, --gaps-only, and --interactive flags to pace execution, focus on gap closure, or run sequential interactive checkpoints for pair-programming style verification.
  • Verification & State Management: Preserve workflow gates, perform phase verification, and update routing/state only when plans are complete.
  • Use Case: Run a phased feature rollout where dependent tasks must be executed in controlled waves to manage quotas and enable staged verification.

Quick Start

Execute gsd-execute-phase for phase 3 with --wave 1 to run only the first wave of plans and collect subagent results for verification.

Frequently Asked Questions about gsd-execute-phase

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I execute project plans with wave-based parallelization?

Wave-based parallelization executes project plans by analyzing task dependencies, grouping them into staged waves, and running isolated subagents in parallel or paced sequences. This removes manual coordination overhead when many dependent tasks must run concurrently.

How does dependency analysis work for grouping tasks into execution waves?

Dependency analysis for execution waves works by discovering plans within a project phase, mapping their dependencies, and grouping dependent tasks into sequential waves. Tasks without dependencies execute in parallel within the same wave, while dependent tasks wait for preceding waves to complete.

Can I run specific execution waves or focus on closing gaps in a project phase?

You can run specific execution waves or focus on closing gaps using the --wave flag to execute only a targeted wave and the --gaps-only flag to concentrate solely on gap closure. These flexible filters pace execution and manage staged rollouts or quotas.

Does interactive mode support pair-programming style verification during task execution?

Interactive mode supports pair-programming style verification by using the --interactive flag to run sequential interactive checkpoints. This allows you to verify subagent results step-by-step and preserve workflow gates before updating routing or project state.

Why spawn isolated subagents with fresh context for plan execution?

Spawning isolated subagents with fresh context ensures deterministic plan handling and result collection. Each subagent loads a clean execution environment, preventing state contamination across parallel tasks and maintaining reliable verification within each wave.

What are the limitations of using wave-based orchestration for phased feature rollouts?

A limitation of wave-based orchestration for phased rollouts is that strict dependency graphs force sequential execution between waves, potentially slowing throughput. Additionally, routing and state updates only occur when all plans complete, requiring full wave verification before proceeding.