dev-team

Orchestrate multi-agent teams to classify, investigate, plan, and execute complex tasks with quality gates.

Updated May 5, 2026

Search Tags:#code review #multi-agent #quality gates #agent coordination #evidence-based #strategy planning #team orchestration

Ready to run

npx skills add https://github.com/swghosh/ambient-workflows --skill dev-team-swghosh

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: dev-team
Source: https://github.com/swghosh/ambient-workflows/tree/main/workflows/dev-team/.claude/skills/dev-team
Command: npx skills add https://github.com/swghosh/ambient-workflows --skill dev-team-swghosh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill solves the problem of getting complex tasks done reliably by orchestrating the right mix of AI agents with quality gates, clear scopes, and phased execution.

Core Features & Use Cases

Task classification & deep investigation: Determines whether you need code changes, PR review, doc review, strategy, or communication, then gathers evidence before acting.
Team design with quality gates: Assembles 2–4 role-specific agents (including a Checker) with defined responsibilities and standards.
Plan-first execution for risky work: Uses approval workflows for complex/unfamiliar implementation to reduce rework and unsafe changes.
Evidence-based output delivery: Requires file/line evidence for claims and enforces actionable, confidence-based review feedback.

Quick Start

Tell the AI: “Use dev-team to investigate and propose a safe implementation plan for improving our authentication flow, then produce a verified PR-ready report.”

Frequently Asked Questions about dev-team

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

FAQPage Schema

How do I coordinate multiple AI agents for complex software engineering tasks?▼

Coordinating multiple AI agents involves assembling a 2-4 role-specific team, including a Checker, to handle phased execution with enforced quality gates for software engineering work. This approach ensures reliable results through role-based task ownership and evidence-backed outputs.

Can I enforce code review quality gates with file and line citations?▼

Enforcing code review quality gates requires evidence-based output delivery that demands file and line citations for all claims. This ensures actionable, confidence-based review feedback by applying checker and security reviews where appropriate during the process.

What is the best way to plan a safe implementation strategy for unfamiliar code changes?▼

Planning a safe implementation strategy uses a plan-first execution workflow that requires approval before acting on complex or unfamiliar code changes. This reduces rework and unsafe modifications by enforcing phased scoping and deep investigation.

Does multi-agent orchestration work for strategy research and communication drafting?▼

Multi-agent orchestration works for strategy research and communication drafting by classifying the task type first, then assigning role-specific agents to investigate and execute. This ensures evidence-backed outputs tailored to the specific requirements of each domain.

When should I use multi-agent orchestration instead of a single agent for code implementation?▼

Use multi-agent orchestration instead of a single agent for complex code implementation requiring verified results, deep investigation, and enforced quality gates. It is necessary when tasks demand role-based ownership, phased scoping, and evidence-backed outputs with file citations.