skills.rest
erenisiklar
erenisiklarCOMMUNITY
Explore all skills by this author →
GitHub

subagent-driven-development

Dispatch fresh subagents per task and enforce two-stage reviews.

Updated Mar 27, 2026
Search Tags:#automation#workflow#planning#tdd#development#subagent#code-review
One-click install
npx skills add https://github.com/erenisiklar/Pusula --skill subagent-driven-development-erenisiklar
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/erenisiklar/Pusula/tree/main/.claude/skills/subagent-driven-development
Command: npx skills add https://github.com/erenisiklar/Pusula --skill subagent-driven-development-erenisiklar

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Subagent-Driven Development addresses the challenge of coordinating complex implementation plans by dispatching a fresh subagent per task, ensuring isolated context and fast feedback loops.

Core Features & Use Cases

  • Fresh subagents per task to keep context clean and reduce cross-task contamination.
  • Two-stage review after each task: spec compliance review, followed by code quality review.
  • Per-task orchestration within the same session or across parallel sessions.
  • Self-checks and structured reporting to enable rapid iteration and safe handoffs.

Quick Start

Dispatch an implementer subagent for Task N with the full task text and context, then route the work through the spec-compliance reviewer and the code-quality reviewer in sequence.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I automate plan execution using subagents for isolated context?▼

Automate plan execution by dispatching a fresh subagent per task to ensure isolated context and prevent cross-task contamination. This approach keeps context clean and enables fast feedback loops during sequential plan implementation.

What is the two-stage review process in subagent-driven development?▼

The two-stage review process verifies each task sequentially through a spec-compliance reviewer followed by a code-quality reviewer. This enforces specification adherence and code standards before completing any dispatched task.

Do I need a predefined plan to use subagent-driven development?▼

Yes, you need a defined sequential plan with per-task prompts for the implementer, spec-reviewer, and code-quality-reviewer, plus reviewer templates. The workflow orchestrates independent tasks within the same session.

Can I execute independent tasks in parallel sessions with subagent orchestration?▼

Yes, subagent orchestration supports per-task execution within the same session or across parallel sessions. Each task receives per-task isolation to prevent context bleed and ensure clean independent handoffs.

Why does context bleed between sequential development tasks?▼

Context bleeds when sequential tasks share the same execution environment without isolation. Dispatching fresh subagents per task prevents this cross-task contamination and maintains clean context for rapid feedback.

What's the best way to enforce spec compliance and code quality in automated workflows?▼

Enforce spec compliance and code quality by routing automated workflow output through two sequential reviewer stages. Spec-reviewers verify specification compliance while code-quality-reviewers check standards using structured templates.

Related Skills

Explore More Software Engineering Skills→
taku-otaku-o
1

kiro-impl

Implement approved development tasks with TDD cycles and subagent reviewers.

Community
Advanced
KYRIE66nbKYRIE66nb

subagent-driven-development

Orchestrates subagents to execute implementation plans task-by-task with two-stage spec and quality reviews.

Community
Advanced
fiatkongenfiatkongen

react-implementer-prompt

Dispatch React frontend subagents with TDD and Tailwind v4 workflows.

Community
Intermediate
synqingsynqing
1

subagent-driven-development

Dispatch fresh subagents per task for isolated implementation and code review.

Community
Advanced