Bundles Deep Dive
This page provides detailed information about every bundle in the Amplifier ecosystem. The data is generated from
data/catalog.json, which is scanned from actual bundle repositories.
How to read this
Each bundle shows: what it includes, which modules it provides, version, and where it lives. This helps you understand
what you're getting when you use a bundle.
All Bundles (scanned from ecosystem)
Understanding Bundle Details
Bundle Fields
- Name - How you reference the bundle (
amplifier bundle use name) - Version - Semantic version (helps track changes)
- Description - What the bundle does
- Includes - Other bundles this one builds on
- Repository - Where the bundle source lives
Module Types in Bundles
- Providers - AI model backends (Anthropic, OpenAI, etc.)
- Tools - Capabilities for the agent (filesystem, bash, web, etc.)
- Hooks - Lifecycle behaviors (logging, approvals, notifications, etc.)
- Orchestrator - How the agent loop runs (basic, streaming, events)
- Context - How conversation state is managed
Bundle Categories
Foundation Bundles
Core bundles that provide base functionality. Most custom bundles include foundation.
Domain Bundles
Specialized for specific use cases: code intelligence (LSP), design work (design-intelligence), workflow orchestration (recipes), issue tracking (issues).
Behavior Bundles
Small, composable bundles that add specific hooks or behaviors. Often included by other bundles rather than used alone.
How to Use This Information
Choosing a bundle
- Start with
foundationfor general use - Use domain bundles for specialized work (design, recipes, LSP)
- Check what each bundle includes to avoid redundancy
- Look at version numbers to assess maturity
Creating your own bundle
Study bundles similar to what you want to build:
- Need base capabilities? Look at
foundation - Extending another bundle? Look at
lsp-pythonpattern - Adding behaviors? Look at notification/logging bundles