Agent Brain Plugin Quickstart

Quickstart: Agent Brain Plugin

Feature: 114-agent-brain-plugin Date: 2025-01-31

Overview

This feature converts the existing Agent Brain skill into a full Claude Code plugin with commands, agents, and skills for document search and setup assistance.

Prerequisites

Before working on this feature:

1. Understand the plugin structure:

   • Read data-model.md for entity definitions
   • Read contracts/command-schema.md for command format
   • Read contracts/agent-schema.md for agent format
   • Read contracts/skill-schema.md for skill format

2. Understand the source material:

   • Existing skill: agent-brain-skill/using-agent-brain/
   • CLI commands: agent-brain-cli/
   • Plugin template: ~/.claude/skills/creating-plugin-from-skill/

3. Required skills to reference:

   • creating-plugin-from-skill - Plugin creation workflow
   • skill-creator - Skill best practices
   • improving-skills - Quality grading rubric

Quick Reference

Plugin Structure

agent-brain-plugin/
├── .claude-plugin/
│   └── marketplace.json      # Plugin manifest
├── commands/                  # 15 slash commands
│   ├── agent-brain-search.md
│   ├── agent-brain-semantic.md
│   ├── agent-brain-keyword.md
│   ├── agent-brain-install.md
│   ├── agent-brain-setup.md
│   ├── agent-brain-config.md
│   ├── agent-brain-init.md
│   ├── agent-brain-verify.md
│   ├── agent-brain-start.md
│   ├── agent-brain-stop.md
│   ├── agent-brain-status.md
│   ├── agent-brain-list.md
│   ├── agent-brain-index.md
│   ├── agent-brain-reset.md
│   └── agent-brain-help.md
├── agents/                    # 2 proactive agents
│   ├── search-assistant.md
│   └── setup-assistant.md
├── skills/                    # 2 skills with references
│   ├── using-agent-brain/
│   │   ├── SKILL.md
│   │   └── references/
│   │       ├── hybrid-search-guide.md
│   │       ├── bm25-search-guide.md
│   │       ├── vector-search-guide.md
│   │       └── api_reference.md
│   └── agent-brain-setup/
│       ├── SKILL.md
│       └── references/
│           ├── installation-guide.md
│           ├── configuration-guide.md
│           └── troubleshooting-guide.md
├── README.md
└── .gitignore

Command Categories

Category	Commands	Purpose
Search	/agent-brain-search, /agent-brain-semantic, /agent-brain-keyword	Document search
Setup	/agent-brain-install, /agent-brain-setup, /agent-brain-config, /agent-brain-init, /agent-brain-verify	Installation & config
Server	/agent-brain-start, /agent-brain-stop, /agent-brain-status, /agent-brain-list	Server management
Indexing	/agent-brain-index, /agent-brain-reset	Document indexing
Help	/agent-brain-help	Usage help

CLI Command Mapping

Plugin Command	CLI Command
/agent-brain-search "query"	agent-brain query "query" --mode hybrid
/agent-brain-semantic "query"	agent-brain query "query" --mode vector
/agent-brain-keyword "query"	agent-brain query "query" --mode bm25
/agent-brain-start	agent-brain start --daemon
/agent-brain-status	agent-brain status
/agent-brain-index /path	agent-brain index /path

Development Workflow

Phase 1: Foundation

1. Create output directory: agent-brain-plugin/
2. Create marketplace.json with all references
3. Create README.md and .gitignore

Phase 2: Commands

1. Create each command file in commands/
2. Use contract schemas from contracts/command-schema.md
3. Include CLI mapping, output formatting, error handling

Phase 3: Agents

1. Create agent files in agents/
2. Use contract schemas from contracts/agent-schema.md
3. Test trigger patterns match intended use cases

Phase 4: Skills

1. Create skill directories in skills/
2. Migrate content from existing agent-brain-skill/
3. Create reference files with detailed documentation

Phase 5: Quality

1. Grade skills using improving-skills rubric
2. Target 95% score for both skills
3. Iterate until quality threshold met

Phase 6: Publish

1. Create GitHub repository
2. Push plugin to SpillwaveSolutions/agent-brain-plugin
3. Document installation instructions

Quality Checklist

Before completing each phase:

• Files follow contract schemas exactly
• YAML frontmatter validates correctly
• Markdown body includes all required sections
• Links between files are correct
• No duplicate or conflicting definitions

Common Issues

Command Parameters

• Use lowercase, hyphenated parameter names
• Mark required parameters explicitly
• Provide sensible defaults for optional parameters

Agent Triggers

• Test regex patterns for false positives
• Ensure patterns cover common phrasings
• Avoid overly broad patterns that trigger too often

Skill References

• Keep SKILL.md under 500 lines
• Put details in references/ folder
• Use descriptive file names for references

Testing

After implementation:

1. Install plugin: cd ~/.claude/plugins && git clone <url>
2. Restart Claude Code
3. Test each command works correctly
4. Verify agents trigger on appropriate messages
5. Confirm skills load when referenced

Resources

• Creating Plugin from Skill
• Skill Creator
• Improving Skills
• Agent Brain CLI
• Existing Skill