skills.rest
Cloud-Byte-Consulting
Cloud-Byte-ConsultingOFFICIAL
Explore all skills by this author →
GitHub

agent-extension-decision-tree

Classifies workflows into prompt, skill, plugin, or MCP packaging paths.

Updated Jul 16, 2026
Search Tags:#decision tree#skill#plugin packaging#agent extension#mcp integration#workflow classification
One-click install
npx skills add https://github.com/Cloud-Byte-Consulting/plugins --skill agent-extension-decision-tree-cloud-byte-consulting
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-extension-decision-tree
Source: https://github.com/Cloud-Byte-Consulting/plugins/tree/main/prompt-workflows/skills/agent-extension-decision-tree
Command: npx skills add https://github.com/Cloud-Byte-Consulting/plugins --skill agent-extension-decision-tree-cloud-byte-consulting

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When you have a reusable workflow, it is hard to know whether it should be packaged as a plain prompt, a skill, a plugin, or a plugin with MCP integrations. This Skill guides you through a structured decision process to pick the smallest extension surface that fits your workflow's complexity. ## Core Features & Use Cases - Structured Intake: Asks targeted clarifying questions about workflow frequency, repeatability, external system dependencies, and asset needs. - Six-Path Classification: Classifies the workflow into exactly one of six build paths, from plain prompt to plugin with MCP server or app integration. - Workflow Audit Support: Accepts an existing Workflow Audit as input to speed up the decision. - Use Case: You have a weekly reporting routine that pulls data from GitHub and Slack. Describe it to the Skill, answer its questions, and receive a recommendation on whether it should remain a prompt or become a plugin with integrations. ## Quick Start Describe the workflow you want to package and ask the agent extension decision tree to recommend the right packaging level.

Frequently Asked Questions about agent-extension-decision-tree

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

FAQPage Schema
How do I decide between a prompt, skill, or plugin for my workflow?▼

Describe the workflow and answer clarifying questions about frequency, repeatability, external integrations, and assets. The Skill classifies it into exactly one of six build paths, steering you to the simplest option that solves the problem.

When should a workflow become a plugin with an MCP server?▼

A workflow needs a plugin with MCP integration when it depends on external systems at runtime, such as GitHub, Slack, Drive, Figma, databases, or a browser, and cannot operate from pasted context and local files alone.

What questions does the decision tree ask about my workflow?▼

It asks how often the workflow runs, whether it follows the same steps each time, whether it needs external systems, whether others must install it, whether it has multiple phases, and whether it depends on templates, scripts, or integrations.

Can I use a Workflow Audit as input to the decision tree?▼

Yes. If you already have a Workflow Audit, you can paste it in when asked to describe the workflow, which gives the classification process a head start.

When is a plain prompt enough instead of a skill or plugin?▼

A plain prompt is sufficient when the workflow runs infrequently, works from pasted context and local files, has no external runtime dependencies, and only you need to use it.

Related Skills

Explore More Software Engineering Skills→
bearhedgebearhedge

writing-plans

Generate step-by-step Markdown implementation plans from feature specs.

Community
Basic
christophevgchristophevg

plugin-development

Scaffold Claude Code plugin structures with skills, agents, hooks, and MCP servers.

Community
Intermediate
VibeTechnologiesVibeTechnologies
1

github-apps

Automates provisioning, credential rotation, and installation validation of GitHub Apps for VibeTeam roles.

Official
Advanced
liveview-nativeliveview-native
1

oneshot

Execute tasks end-to-end with retries and a final summary.

Official
Advanced