subagent-driven-development

Subagent-Driven Development

Overview

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

Core principle: Fresh subagent per task + two-stage review (spec then quality) = high quality, fast iteration.

When to Use

Use this skill when:

• You have an implementation plan (from writing-plans skill or user requirements)
• Tasks are mostly independent
• Quality and spec compliance are important
• You want automated review between tasks

vs. manual execution:

• Fresh context per task (no confusion from accumulated state)
• Automated review process catches issues early
• Consistent quality checks across all tasks
• Subagents can ask questions before starting work

The Process

1. Read, Parse, and Preflight the Plan

Read the plan file. Before dispatching implementation, audit it for internal contradictions and claims that cannot simultaneously hold—especially low-latency streaming versus authoritative final-channel precedence, bounded shutdown versus retained worker ownership, retry claims versus partial-stream replay safety, and install/public-link claims for artifacts not yet published. Correct the execution contract or explicitly document the tradeoff before production code begins.

For plans with streamed irreversible side effects (TTS, media, notifications, progressive writes), load during preflight.

Extract ALL tasks with their full text and context upfront. Create a todo list:

# Read the plan
read_file("docs/plans/feature-plan.md")

# Create todo list with all tasks
todo([
 {"id": "task-1", "content": "Create User model with email field", "status": "pending"},
 {"id": "task-2", "content": "Add password hashing utility", "status": "pending"},
 {"id": "task-3", "content": "Create login endpoint", "status": "pending"},
])

Key: Read the plan ONCE. Extract everything. Don't make subagents read the plan file — provide the full task text directly in context.

2. Worker Transaction Contract