skills.rest
epicodic
epicodicCOMMUNITY
Explore all skills by this author →
GitHub

subagent-driven-development

Executes implementation plans by dispatching fresh subagents per task with ordered review gates.

Updated Aug 28, 2026
Search Tags:#workflow#tdd#code review#implementation plan#subagent#spec compliance#task orchestration
One-click install
npx skills add https://github.com/epicodic/drift --skill subagent-driven-development-epicodic
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/epicodic/drift/tree/main/.agents/skills/subagent-driven-development
Command: npx skills add https://github.com/epicodic/drift --skill subagent-driven-development-epicodic

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Executing a multi-task implementation plan manually risks context pollution, skipped reviews, and inconsistent quality. This Skill orchestrates plan execution by dispatching a fresh subagent per task and enforcing ordered quality gates so every task is verified before moving on. ## Core Features & Use Cases - Fresh Subagent Per Task: Each task runs in an isolated subagent context with curated instructions, preventing context pollution and preserving the controller's coordination capacity. - Ordered Quality Gates: Every task passes spec compliance review, code quality review, and a mandatory coding-guideline checklist before being marked complete. - Status Handling and Model Selection: Implementer statuses (DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, BLOCKED) drive escalation logic, and model tiers are matched to task complexity. - Use Case: Given a written plan with five independent tasks, the controller extracts all tasks, dispatches implementer and reviewer subagents in sequence, loops on review findings, and finishes with a final whole-implementation review. ## Quick Start Ask the AI to execute the implementation plan at docs/agents/plans/feature-plan.md using subagent-driven development with per-task spec and code quality reviews.

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 AI subagents?▼

Extract all tasks from the plan into a todo list, then dispatch a fresh implementer subagent per task with the full task text and context. After each task, run spec compliance and code quality reviewer subagents before marking it complete.

What is the difference between subagent-driven development and executing plans in a parallel session?▼

Subagent-driven development stays in the same session with fresh subagents per task and automatic review checkpoints, enabling faster iteration. Executing plans uses a separate parallel session with a handoff, which suits work that should not share the current session.

When should I not use subagent-driven development?▼

Avoid it when tasks are tightly coupled and cannot be implemented independently, when no implementation plan exists yet, or when you cannot dispatch subagents in the current session. Brainstorm or write a plan first in those cases.

How do I handle a blocked or failing implementer subagent?▼

Read the reported status: provide missing context for NEEDS_CONTEXT, and for BLOCKED either add context, re-dispatch with a more capable model, split the task, or escalate to the human. Never force the same model to retry unchanged.

Which model should I use for each subagent task?▼

Match model capability to task complexity: cheap fast models for mechanical one-to-two-file tasks with clear specs, standard models for multi-file integration work, and the most capable model for architecture, design, and review tasks.

Related Skills

Explore More Software Engineering Skills→
HumanBean17HumanBean17
8

subagent-driven-development

Dispatch isolated subagents to implement and review tasks from a structured plan file.

Community
Advanced
EtanHeyEtanHey

pr-loop

Orchestrate the complete Pull Request lifecycle from branching to merge and cleanup.

Community
Advanced
ZxelaZxela

implement

Implement acceptance criteria through a test-driven development workflow.

Community
Advanced
ShunmeiChoShunmeiCho
2

subagent-driven-development

Dispatch fresh subagents per task with built-in code-review gates.

Community
Advanced