subagent-driven-development

Execute implementation plans by dispatching fresh subagents with two-stage review.

1.4k|129|Updated Mar 13, 2026
One-click install
npx skills add https://github.com/shiwenwen/hope-agent --skill subagent-driven-development-shiwenwen
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/shiwenwen/hope-agent/tree/main/skills/subagent-driven-development
Command: npx skills add https://github.com/shiwenwen/hope-agent --skill subagent-driven-development-shiwenwen

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill solves the problem of implementing an execution plan reliably by coordinating task decomposition, fresh execution context per task, and automated gatekeeping to prevent spec drift and low-quality code.

Core Features & Use Cases

  • Fresh subagent per task: Dispatches a clean implementer for each task to reduce context pollution and cross-task confusion.
  • Two-stage review per task: Runs a spec-compliance review first, then a code-quality review to catch correctness and workmanship issues early.
  • Plan-driven task execution: Reads and parses an implementation plan once, extracts the full task specs, and then executes tasks using those concrete instructions.
  • Use Case: When you have an implementation plan for a feature (e.g., adding an auth endpoint and data model changes), this Skill executes each task with TDD-aligned instructions and iterative review loops before moving on.

Quick Start

Use the subagent-driven-development skill to execute an existing implementation plan by dispatching an implementer and running spec-compliance and code-quality reviews for each extracted task.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I execute an implementation plan with automated spec compliance and code quality reviews?

To execute an implementation plan with automated reviews, dispatch a fresh subagent per task and run two-stage gated reviews. This enforces spec compliance before code quality review, iterating rework until both gates approve.

What is the best way to prevent context pollution when executing multiple independent coding tasks?

The best way to prevent context pollution across independent coding tasks is dispatching a fresh subagent for each task. This isolates execution context, preventing cross-task confusion and ensuring accurate task implementation.

How does TDD-oriented development work with subagent task orchestration?

TDD-oriented development with subagent task orchestration works by injecting full task text into a clean implementer subagent. It enforces test-driven instructions and iterative review loops before moving to the next task.

Can I use subagent-driven development for features requiring commit readiness checks?

Yes, you can use subagent-driven development for features requiring commit readiness checks. It enforces correctness and code quality through two-stage reviews before allowing any changes to be committed.

When do I need two-stage gated reviews for feature implementation?

You need two-stage gated reviews for feature implementation when tasks are mostly independent and correctness must be strictly enforced. Running spec-compliance before code-quality review catches drift and workmanship issues early.

Why does my implementation plan execution drift from the original spec during coding?

Implementation plan execution drifts from the original spec when context pollution accumulates across tasks. Dispatching fresh subagents per task and enforcing spec-compliance review before quality review prevents this drift.