subagent-driven-development

Delegate each task to a fresh subagent with isolated context.

2|Updated Mar 3, 2026

Search Tags:#automation #isolation #subagents #prompt-engineering #quality-review #task-decomposition #ai-workflow

Ready to run

npx skills add https://github.com/MomoDaviluke/star-citizen-promotion --skill subagent-driven-development-momodaviluke

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/MomoDaviluke/star-citizen-promotion/tree/main/.trae/skills/subagent-driven-development
Command: npx skills add https://github.com/MomoDaviluke/star-citizen-promotion --skill subagent-driven-development-momodaviluke

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates complex plans by delegating each task to a fresh subagent with an isolated context, preventing cross-task contamination and enabling rapid iteration.

Core Features & Use Cases

Per-task subagents with isolated contexts to execute plans without session bleed.
Two-stage reviews: first verify specification compliance, then assess code quality.
Clear task lifecycle and status handling (DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, BLOCKED) to manage progress.
Reusable prompts and templates for implementers and reviewers to ensure consistent quality.
Safe, auditable collaboration workflow that scales with task count.

Quick Start

Define your plan and deploy per-task subagents to execute and review each task.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema

How does subagent-driven development prevent cross-task context contamination?▼

Subagent-driven development prevents context contamination by delegating each task to a fresh subagent with an isolated context, ensuring no session bleed occurs during multi-step plan execution. This enables rapid iteration without prior tasks influencing current outcomes.

How do I automate multi-step planning workflows with isolated subagents?▼

You automate multi-step planning workflows by defining a plan and deploying per-task subagents using implementer-prompt.md, spec-reviewer-prompt.md, and code-quality-reviewer-prompt.md. Each subagent executes independently with statuses and guards for completion tracking.

What task lifecycle statuses are used for managing subagent execution progress?▼

Task lifecycle statuses for managing subagent execution progress include DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, and BLOCKED. These statuses provide clear guards for completion and enable safe, auditable collaboration workflows that scale with task count.

When should I use isolated subagents for task decomposition instead of a single context?▼

Use isolated subagents for task decomposition when tasks can be independently executed but require rigorous specification verification and code quality review. This approach is ideal for complex plans where preventing cross-task contamination and enabling rapid iteration are critical.

Do I need predefined prompts to start automating task execution with subagents?▼

Yes, automating task execution with subagents requires per-task prompts and a defined workflow using implementer-prompt.md, spec-reviewer-prompt.md, and code-quality-reviewer-prompt.md. These reusable templates ensure consistent quality and safe, auditable collaboration.