Claude Agent Skills & Plugins: Automation Workflows (Part 3)

Welcome to the final installment of the SEOGEO Claude Mastery Series. You already know how to run commands and give Claude external MCP tools. Now, we unlock true autonomy.

By defining Agent Skills and leveraging Plugins, you transition Claude from a simple chatbot into a model-invoked teammate that actively reviews your code, standardizes your commits, and manages complex multi-step workflows without constant human prompting.

🧠 1. Agent Skills - Modular Capabilities

What are Agent Skills?
Skills are modular capabilities that extend Claude's core functionality. The key difference between a Skill and a Slash Command is autonomy:

• Slash Commands: User-invoked (you type /command). Best for simple, single-file prompts.
• Agent Skills: Model-invoked. Claude autonomously decides to use them based on the context of your conversation. Best for comprehensive workflows spanning multiple files and scripts.

Skill Setup Commands

Skills are automatically discovered by Claude based on their location and YAML description. Here is how to create the directory structure:

Create a project-specific Skill directory:

mkdir -p .claude/skills/skill-name

Create a global/personal Skill directory:

mkdir -p ~/.claude/skills/skill-name

Create the Skill definition file:

cat > .claude/skills/my-skill/SKILL.md

📄 2. Skill Templates & Examples

A Skill is defined by a SKILL.md file that contains YAML frontmatter (for Claude's system parser) and Markdown instructions (for execution).

Template A: Simple Skill (Commit Helper)

File: .claude/skills/commit-helper/SKILL.md

---
name: generating-commit-messages
description: Generates clear commit messages from git diffs. Use when writing commit messages or reviewing staged changes.
---

# Generating Commit Messages

## Instructions

1. Run `git diff --staged` to see changes
2. I'll suggest a commit message with:
   - Summary under 50 characters
   - Detailed description
   - Affected components

## Best practices

- Use present tense
- Explain what and why, not how

Template B: Skill with Tool Permissions

If you want Claude to perform a task but want to restrict it from running destructive terminal commands, use the allowed-tools key in your YAML.

File: .claude/skills/code-reviewer/SKILL.md

---
name: code-reviewer
description: Review code for best practices and potential issues. Use when reviewing code, checking PRs, or analyzing code quality.
allowed-tools: Read, Grep, Glob
---

# Code Reviewer

## Review checklist

1. Code organization and structure
2. Error handling
3. Performance considerations
4. Security concerns
5. Test coverage

🧩 3. Plugins - Extend Claude Code

What are Plugins?
Plugins are distributable packages that extend Claude Code. A single plugin can provide Custom Commands, Specialized Agents, Agent Skills, Automation Hooks (event triggers), and pre-configured MCP Servers. You can install them from marketplaces or create your own.

Plugin Management Commands

Open plugin interface UI:

/plugin

Add a plugin marketplace:

/plugin marketplace add <url-or-path>

Install a plugin:

/plugin install <name>@<marketplace>

Enable/Disable/Uninstall Plugins:

/plugin enable <name>@<marketplace>
/plugin disable <name>@<marketplace>
/plugin uninstall <name>@<marketplace>

Quick Start Marketplace Examples

Add a GitHub Marketplace (e.g., your organization's private plugins):

claude mcp add marketplace your-org/claude-plugins

Install a specific plugin from a marketplace:

/plugin install formatter@your-org

Add a Local Development Marketplace:

/plugin marketplace add ./dev-marketplace