03
Docs

Frameworks

Compound Engineering and Compound AI Systems for agentic development

Two complementary frameworks for understanding and applying agentic development.

Compound Engineering

Source: every.to/guides/compound-engineering

Core Principle: Each unit of work should make subsequent work easier, not harder.

The Compound Loop:

Plan → Work → Review → Compound → Repeat

Key Concepts:

ConceptDescription
Verification-firstAlways give AI a way to verify its work (tests, types, linting)
Infrastructure investmentTesting harness + CI/CD = exponential dividends
Mindset shiftCode writer → System orchestrator
Time allocation80% planning + review, 20% code generation

The Maturity Ladder:

  1. Manual development (no AI)
  2. AI as reference (copy-paste snippets)
  3. AI as assistant (inline suggestions)
  4. AI as partner (multi-file editing)
  5. AI as agent (compound engineering) ← Workshop goal

Compound Engineering Plugin

Every.to has released an official Claude Code plugin that implements these principles:

Install:

/plugin marketplace add https://github.com/EveryInc/compound-engineering-plugin
/plugin install compound-engineering

Commands:

CommandPurpose
/workflows:planTurn feature ideas into detailed implementation plans
/workflows:workExecute plans with worktrees and task tracking
/workflows:reviewMulti-agent code review before merging
/workflows:compoundDocument learnings to make future work easier

Why it matters: This plugin codifies the compound engineering workflow — plans inform future plans, reviews catch more issues, patterns get documented.

Similar Plugins Worth Exploring

Note

These plugins are optional and not used in the workshop labs. Explore them after the workshop as you develop your own workflow.

PluginPurposeInstall
Ralph WiggumAutonomous iteration loops — Claude works repeatedly until completion/plugin install ralph-wiggum
Context7Live documentation lookup — reduces hallucinations with current API docs/plugin install context7
Feature DevelopmentStructured workflow: explore → design → implement → review/plugin install feature-development
Security GuidanceReal-time security scanning during code generation/plugin install security-guidance

Compound AI Systems

Source: Berkeley AI Research

Core Principle: State-of-the-art AI results come from systems with multiple components, not monolithic models.

Architecture:

Key Patterns:

PatternComponentsExample
RAGLLM + retriever + vector DBDocumentation Q&A
AgentLLM + tools + executorClaude Code
Multi-AgentMultiple LLMs + coordinatorResearch + Writing agents
HybridNeural + symbolicCode generation + linting

Why It Matters:

How the Frameworks Apply

FrameworkAnswersWorkshop Application
Compound EngineeringHow do we work with AI?Planning, verification, iteration patterns
Compound AI SystemsWhat do we build with AI?Multi-component architectures, MCP pipelines