subagent-driven-development

Execute implementation plans by delegating tasks to fresh subagents with two-stage review.

Updated May 24, 2026
One-click install
npx skills add https://github.com/anxiety135790/hermes-skills --skill subagent-driven-development-anxiety135790
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/anxiety135790/hermes-skills/tree/main/software-development/subagent-driven-development
Command: npx skills add https://github.com/anxiety135790/hermes-skills --skill subagent-driven-development-anxiety135790

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you execute implementation plans reliably by delegating each task to a fresh subagent and enforcing review checkpoints, reducing context pollution and catching mistakes early.

Core Features & Use Cases

  • Plan-driven execution: Read a plan once, extract all tasks, and convert them into a structured todo list before work begins.
  • Fresh subagent per task: Assign each task to a new implementer subagent with complete context, including explicit TDD instructions and task-specific constraints.
  • Two-stage review: Verify every task first for spec compliance and then for code quality before moving on.
  • Final integration review: Perform a last pass over the full implementation to catch cross-task inconsistencies and merge risks.
  • Use case: A developer with a feature plan can hand off each small implementation step to subagents, ensure each change meets the spec, and keep the whole workflow consistent and testable.

Quick Start

Use the subagent-driven-development skill to execute my implementation plan by delegating one task at a time with spec review and quality review after each 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 using subagents without losing context?

To execute an implementation plan without context pollution, this Skill delegates each task to a fresh subagent. It extracts tasks into a structured todo list and provides complete context, ensuring isolated execution and reducing context loss.

What is the best way to enforce code review and TDD when delegating feature delivery to subagents?

The best way to enforce TDD and code review during feature delivery is using a two-stage review process. This Skill mandates spec compliance and code quality checks after each fresh subagent completes its assigned implementation task.

Can I use plan-based execution for independent tasks with this subagent workflow?

Yes, you can use this subagent workflow for plan-based execution of independent tasks. It reads your plan, extracts all tasks into a todo list, and assigns each implementation step to a new subagent with explicit TDD instructions.

How do you catch cross-task inconsistencies and merge risks after delegating implementation tasks?

To catch cross-task inconsistencies and merge risks, this workflow performs a final integration review. It executes a last pass over the full implementation to verify consistency across all delegated tasks before final delivery.

Why should I use a fresh subagent per task instead of a single agent for software engineering workflows?

Using a fresh subagent per task reduces context pollution and catches mistakes early. This approach enforces review checkpoints for spec compliance and code quality after each isolated task, unlike single-agent execution.