skills.rest
stabilefrisur
stabilefrisurCOMMUNITY
Explore all skills by this author →
GitHub

subagent-driven-development

Dispatch fresh subagents per task with two-stage spec and quality reviews.

Updated Mar 19, 2026
Search Tags:#automation#subagent#code-review#autonomous#task-orchestration#development-workflow#AI-assist
One-click install
npx skills add https://github.com/stabilefrisur/panmetis --skill subagent-driven-development-stabilefrisur
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/stabilefrisur/panmetis/tree/main/src/panmetis/skills/subagent-driven-development
Command: npx skills add https://github.com/stabilefrisur/panmetis --skill subagent-driven-development-stabilefrisur

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a structured approach to executing implementation plans by spawning independent subagents per task and performing two-stage reviews after each task to ensure specification compliance and code quality.

Core Features & Use Cases

  • Fresh subagent per task for isolated context, fast iteration, and reduced cross-task context leakage.
  • Per-task implementation with tests, self-review, and staged reviews (spec compliance then code quality).
  • Use Case: In a multi-task plan, dispatch subagents for each task, collect results, and merge into a final release.

Quick Start

Instruct an implementer subagent to begin the first task and initiate spec and 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 orchestrate multiple development tasks with isolated contexts using AI subagents?▼

To orchestrate tasks with isolated contexts, dispatch a fresh autonomous subagent per task. This approach ensures fast iteration and prevents cross-task context leakage during multi-task plan execution.

What is the best way to automate code reviews for independent implementation tasks?▼

Automating code reviews for independent tasks is best handled through a two-stage review process. After each subagent completes implementation, staged reviews verify specification compliance and then evaluate overall code quality.

How do I execute a multi-task development plan without cross-task context leakage?▼

Executing a multi-task plan without context leakage involves spawning independent subagents for each task. This isolation allows rapid iteration per task, followed by formal completion signaling and staged reviews.

Why use autonomous subagents for task orchestration instead of a single continuous development workflow?▼

Autonomous subagents are used for task orchestration to isolate contexts and reduce cross-task leakage. Unlike a continuous workflow, dispatching fresh subagents per task enables faster iteration and targeted two-stage reviews.

Does the subagent-driven development workflow require external dependencies to run?▼

The subagent-driven development workflow requires no external dependencies to run. It operates autonomously by extracting plans, dispatching subagents, and performing self-reviews within the existing environment.

When should I avoid using fresh subagents for task orchestration in software engineering?▼

You should avoid using fresh subagents for task orchestration when tasks are heavily interdependent rather than independent. The workflow is designed for isolated contexts, so tightly coupled tasks may suffer from the separation.

Related Skills

Explore More Software Engineering Skills→
Thiago-Cruz-engThiago-Cruz-eng

subagent-driven-development

Coordinates subagent-driven task execution with staged compliance and code quality reviews.

Community
Advanced
joewinkejoewinke
248

jat-complete

Complete the current JAT task with verification, commit, and signaling.

Community
Advanced
musical-basicsmusical-basics

subagent-driven-development

Dispatch subagents to implement tasks with two-stage spec and code quality reviews.

Community
Advanced
lowmasonlowmason

subagent-driven-development

Dispatch fresh subagents per task with per-task reviews and a final aggregation pass.

Community
Advanced