Skip to main content

​
Claude Code Plugin: Smithers Skill

Ghost doc. Real Claude Code plugin at ~/.claude/plugins/smithers/. Registers the smithers skill so Claude Code can create and run workflows.

​
plugin.json

{
  "name": "smithers",
  "version": "0.1.0",
  "description": "Build AI agents with Solid.js - Declarative JSX for Claude orchestration",
  "author": "William Cory",
  "license": "MIT",
  "repository": "https://github.com/evmts/smithers",
  "keywords": ["orchestration", "multi-agent", "smithers", "workflow", "ai-agents", "claude", "jsx", "reactive"],
  "skills": ["skills/smithers"]
}

​
SKILL.md

---
name: smithers-orchestrator
description: Create and monitor multi-agent AI orchestrations using Smithers framework. Use when user wants to coordinate multiple AI agents, create complex workflows, build agent pipelines, or mentions "smithers", "multi-agent", "orchestration", "workflow", "agent coordination". Strongly recommends plan mode for orchestrations with 3+ phases.
allowed-tools:
  - Read
  - Write
  - Edit
  - Bash
  - Glob
  - Grep
  - Task
user-invocable: true
recommend-plan-mode: true
---

# Smithers Orchestrator

## When to Use
- Orchestrate multiple AI agents working together
- Create complex multi-phase workflows
- Build agent pipelines with state management

## Quick Start
1. Initialize: `bunx smithers-orchestrator init` (creates `.smithers/main.tsx`)
2. Edit `.smithers/main.tsx` to define your agent workflow
3. Monitor: `bunx smithers-orchestrator monitor`

​
EXAMPLES.md

Five complete workflow examples (see ~/.claude/plugins/smithers/skills/smithers/EXAMPLES.md for full code):
  1. Simple Sequential: Three-phase research/implement/test pipeline
  2. Conditional Branching: Branches based on analysis results
  3. Parallel Execution: Frontend/backend/database agents simultaneously
  4. Error Handling and Retry: Automatic retry with recovery fallback
  5. Data Flow Between Phases: Requirements/design/implement/test with structured data passing

​
REFERENCE.md

Component / UtilityPurpose
<Ralph>Loop controller: remounts children each iteration until no agents render or maxIterations is reached
<Claude>Self-executing agent: runs a Claude agent on mount, calls onFinished/onError callbacks
<Phase>Semantic grouping for orchestration phases, used in monitor output
<Step>Fine-grained grouping within a phase
Zustand StoreState management: drives phase transitions via onFinished callbacks
RalphContextContext object exposing iteration count and completion signalling
serialize()Serializes the orchestration tree to an XML plan string