compose-team

Assemble per-sub-agent context bundles from evaluate-story output for deterministic spawning.

Updated Mar 8, 2026
One-click install
npx skills add https://github.com/petrSimonidesXart/xPmGateway --skill compose-team-petrsimonidesxart
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: compose-team
Source: https://github.com/petrSimonidesXart/xPmGateway/tree/main/.gaai/core/skills/delivery/compose-team
Command: npx skills add https://github.com/petrSimonidesXart/xPmGateway --skill compose-team-petrsimonidesxart

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Assembles per-sub-agent context bundles from evaluate-story output to ensure each sub-agent receives exactly the information it needs, preventing context pollution and unnecessary token usage.

Core Features & Use Cases

  • Builds tier-specific context bundles for Planning, Implementation, QA, and MicroDelivery sub-agents.
  • Resolves memory file paths using the memory index before assembling bundles and includes only relevant artefacts and rules.
  • Works across the full deployment spectrum from micro-delivery to specialist orchestration, ensuring deterministic spawning.

Quick Start

Run evaluate-story, then run compose-team to produce per-sub-agent context bundles ready for spawning.

Frequently Asked Questions about compose-team

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

FAQPage Schema
How do I assemble context bundles for sub-agents to prevent token pollution?

Sub-agent context orchestration works by resolving memory paths via an index file, selecting tier-specific stories and rules, and assembling them into bundles for deterministic spawning without loading full memory.

How do I create deterministic sub-agent context bundles from evaluate-story output?

To create deterministic sub-agent context bundles, process evaluate-story output to build tier-specific bundles for Planning, Implementation, QA, and MicroDelivery agents ready for spawning.

What is the best way to orchestrate sub-agent context across different delivery tiers?

You need an evaluate-story output containing task requirements, and a memory index file like contexts/memory/index.md to resolve memory paths before assembling the bundles.

Do I need a memory index file to resolve paths for sub-agent orchestration?

Yes, you need a memory index file such as contexts/memory/index.md to resolve memory file paths before assembling bundles and including only relevant artefacts and rules for each sub-agent.

How do I spawn sub-agents without loading the full memory into context?

To spawn sub-agents without loading full memory, assemble per-agent context bundles that output specific file lists containing only the required stories, rules, and artefacts for that agent.