gsd-execute-phase

Execute phase plans in parallel waves across subagents.

Updated Jan 28, 2025
One-click install
npx skills add https://github.com/aslaii/dotfiles --skill gsd-execute-phase-aslaii
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-execute-phase
Source: https://github.com/aslaii/dotfiles/tree/main/claude/skills/gsd-execute-phase
Command: npx skills add https://github.com/aslaii/dotfiles --skill gsd-execute-phase-aslaii

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates the execution of all plans within a phase using wave-based parallelism to improve throughput and coordination across subagents.

Core Features & Use Cases

  • Wave-based plan execution: groups plans into waves and executes them in parallel where dependencies allow.
  • Phase-level governance: validates completion, handles retries, and updates routing and state.
  • Subagent isolation: spawns per-wave subagents with full context, collecting results and surfacing phase status.

Quick Start

Provide the phase number to execute and optional flags such as --wave, --gaps-only, or --interactive to tailor the run.

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 plans in parallel using subagents?

To execute plans in parallel, you group them into wave-based stages and route actions to isolated subagents. This orchestration enables staged parallel execution where dependencies allow, improving throughput across the workflow.

What is wave-based parallelism for phase execution?

Wave-based parallelism is an execution approach that groups plans into sequential stages called waves. Each wave runs multiple plans concurrently across isolated subagents, collecting results and verifying completion before advancing to the next wave.

How do I run a specific phase plan with interactive flags?

You run a specific phase plan by providing the phase number and optional flags like --interactive, --wave, or --gaps-only. These flags tailor the execution run, allowing you to control wave grouping and handle specific action routing.

Can I retry failed actions during phase orchestration?

Yes, phase orchestration includes phase-level governance that validates completion and handles retries. It updates routing and state while applying robust flag handling with guardrails to ensure reliable subagent execution and phase closure.

Does wave parallelism isolate subagent contexts during execution?

Yes, subagent isolation spawns per-wave subagents with full context for each execution cycle. The orchestrator collects the results from these isolated subagents to surface the overall phase status and verify completion.

When should I use gaps-only mode for phase execution?

Gaps-only mode is used when you need to execute only the missing or incomplete parts of a phase plan. By providing the --gaps-only flag, the orchestrator targets specific action routing to fill execution gaps without re-running completed waves.