gsd-execute-phase

Execute phase plans by grouping tasks into waves and coordinating subagents.

Updated Aug 15, 2025
One-click install
npx skills add https://github.com/gesmith0606/nfl_data_engineering --skill gsd-execute-phase-gesmith0606
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-execute-phase
Source: https://github.com/gesmith0606/nfl_data_engineering/tree/main/.claude/skills/gsd-execute-phase
Command: npx skills add https://github.com/gesmith0606/nfl_data_engineering --skill gsd-execute-phase-gesmith0606

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill executes all plans within a phase by discovering plans, analyzing dependencies, grouping tasks into waves, and coordinating parallel subagents while preserving verification and state-update gates. It reduces manual orchestration overhead and avoids race conditions by enforcing wave and verification semantics.

Core Features & Use Cases

  • Discovery & Dependency Analysis: Locate plans in a phase and compute dependencies to ensure safe ordering.
  • Wave Partitioning & Selective Execution: Group independent plans into waves and optionally run a single wave for pacing or staged rollouts.
  • Subagent Orchestration & Aggregation: Spawn lightweight subagents to run plans with isolated context and collect results for verification and state updates.
  • Flag-driven Behaviors: Respect explicit flags (--wave, --gaps-only, --interactive) only when they appear in invocation arguments.
  • Use Case: Run a phased deployment or a large verification sweep where thousands of independent plans must be executed reliably and in parallel.

Quick Start

Run gsd-execute-phase for phase 2 with the --wave 1 flag to execute only the first wave and collect 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 parallelize plan execution across waves while preserving verification gates?

You can parallelize plan execution by discovering phase plans, analyzing dependencies, and grouping independent tasks into waves for subagent orchestration. This ensures safe ordering while preserving verification and state-update gates during parallel execution.

What is dependency analysis for wave partitioning in multi-plan phases?

Dependency analysis for wave partitioning involves computing dependencies across discovered plans to group independent tasks together. This allows parallel subagents to execute safely within a wave while enforcing gated verification before proceeding to the next.

How do I execute only a specific wave of plans for a staged rollout?

You can execute a specific wave by passing the --wave flag with a wave number in your invocation arguments. The orchestrator runs only that wave's plans, collects results, and respects verification gates for controlled, staged rollouts.

Can I run phased deployments with thousands of independent plans reliably?

Yes, phased deployments with thousands of plans are supported through subagent orchestration and wave partitioning. The system spawns lightweight subagents with isolated context to run plans in parallel, aggregating results for verification and state updates.

Why do I need subagent orchestration for parallel plan execution?

Subagent orchestration is needed to avoid race conditions and reduce manual overhead during parallel plan execution. It spawns isolated context workers for each plan, collecting results securely for verification and state updates across waves.

What are the limitations of selective wave execution flags?

Selective wave execution flags like --wave, --gaps-only, and --interactive are only honored when they explicitly appear in invocation arguments. Omitting them means the orchestrator proceeds through all waves automatically without selective pacing or interactive pauses.