Shannon

for Claude Code

AI Agent Workflow Orchestration Platform

Create agents, organize them into teams, plan tasks with AI, and monitor execution in real-time. All powered by Claude Code.

Download Shannon View on GitHub
6+
Core Pages
DAG
Visual Editor
RT
Real-Time Logs
AI
Task Planner
Dashboard

Command Center Overview

Real-time metrics, agent performance, task analytics, and activity feeds - all at a glance.

Shannon - Dashboard
Shannon
for Claude Code
Dashboard
Workspaces
Agents
MCP Servers
Teams
Sessions
Dashboard Auto-refreshes every 5m
Projects
3
Agents
8
Teams
4
Sessions
12
3 tasks running
Task Status
Done 70% Active 30%
Sessions
Done 50% Active 32%
Success Rate
87%
Code Review
Tests Passed 24/28
Files Changed 47
Completion Trend
Visual Team Builder

Design Agent
Workflows Visually

Drag and drop agents onto the canvas, connect them to define execution flow. Choose from parallel, sequential, or fully custom strategies.

Drag & Drop Canvas
Powered by ReactFlow for smooth interactions
3 Strategy Presets
Parallel, Sequential, or Custom pipelines
Multi-Model Teams
Combine Opus, Sonnet, and Haiku agents
Team Editor - Backend Squad
Agents
Drag onto canvas
Code Writer
opus
added
Reviewer
sonnet
added
Tester
haiku
DevOps
sonnet
Code Writer
opus
Reviewer
sonnet
Tester
haiku
Sequential
Session Planner - Auth Feature
JWT Authentication
api-project · 5 tasks
AI Planner
Start Working
AI Task Planner
The workspace has been analyzed. Here are the proposed tasks for JWT authentication implementation:
1.
Setup JWT middleware
Code Writer
2.
Implement login endpoint
Code Writer deps: Setup JWT middleware
3.
Add registration endpoint
Code Writer deps: Setup JWT middleware
4.
Protect routes with auth middleware
Backend Team deps: Login, Registration
5.
Write unit tests for auth
Tester deps: All above
Accept & Create Tasks
Re-plan
AI Planner

Let AI Plan
Your Tasks

Describe your goal in natural language. The AI analyzes your workspace structure and proposes an optimal task breakdown with dependencies.

Workspace-Aware
Analyzes your codebase before planning
Auto-Dependencies
Automatically resolves task ordering
Smart Assignment
Maps tasks to agents or teams
Workspace

DAG View + Live Chat

Watch your task dependency graph execute in real-time. Chat with agents, review code changes, and manage workflow - all in one split view.

Workspace - API Migration
API Migration 3/5 tasks Running
Stop
DB Schema
Completed
API Routes
Completed
Auth Layer
Running...
Integration
Pending
Tests
Pending
Chat
Changes 3
Auth Layer · Code Writer (opus)
Code Writer
I'll implement the JWT authentication layer. First, let me set up the middleware...
Writing file: auth/middleware.go
Code Writer
Created the JWT middleware with token validation, claims extraction, and refresh token support. Now setting up the auth handler...
Writing file: auth/handler.go
Agent is working...
Send a message to the agent...
Prompt Editor

Monaco-Powered
Prompt Engineering

A purpose-built code editor for crafting Claude system prompts. Semantic syntax highlighting, intelligent autocomplete, and one-click AI improvement.

Semantic Tag Coloring
6 tag categories: Role, Structure, Context, Tool, Thinking, Artifact
Smart Autocomplete
XML tag snippets, {{variables}}, and auto-closing tags
AI Improve
One-click prompt enhancement via Claude with revert/keep
Agent Editor - System Prompt
System Prompt
Supports XML tags, Markdown, and {{variables}}
AI Improve

Added structured XML sections, clear constraints, chain-of-thought instructions, and output format specification for better consistency.

Revert
Keep
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
<persona>
You are a senior backend engineer specializing
in Go and **distributed systems**.
</persona>
 
<instructions>
IMPORTANT: Always write tests first.
- Use framework: {{framework}}
- Follow **clean architecture** patterns
</instructions>
 
<thinking>
Reason step by step before writing code.
</thinking>
 
S
constraints [Structure]
C
context [Context]
A
code [Artifact]
T
chain-of-thought [Pattern]
Role
Structure
Context
Tool
Thinking
Artifact
Everything You Need

Built for Power Users

A complete toolkit for managing AI agent workflows at scale.

Custom Agents

Create agents with specific models (Opus, Sonnet, Haiku), system prompts, tool permissions, and MCP server connections.

Team Workflows

Organize agents into teams with visual DAG editor. Parallel, sequential, or custom execution strategies.

MCP Servers

Connect to Model Context Protocol servers for extended tool capabilities. JSON editor with health checks and testing.

Sessions & Tasks

Create sessions with task queues, dependency graphs, retry policies, and automatic test execution after completion.

Project Workspaces

Link GitHub repos, configure build/test commands, and manage CLAUDE.md context files for each workspace.

Encrypted Vault

Store API keys and environment variables securely with AES-256 encryption. Injected at runtime, never exposed.

Tech Stack

Built With Modern Tools

Go
Backend
W
Wails v2
R
React 18
TS
TypeScript
Tw
Tailwind CSS
🐻
Zustand
RF
ReactFlow
SQ
SQLite
How It Works

4 Steps to AI-Powered Development

1

Add Workspace

Link your project repository and configure build/test commands

2

Create Agents

Build specialized agents with models, prompts, and tool permissions

3

Build Teams

Organize agents into teams with visual DAG workflows

4

Launch Session

Plan tasks with AI, execute, and monitor everything in real-time

Ready to Orchestrate?

Start managing your AI agent workflows with Shannon. Open source, local-first, and built for developers.

Download Shannon Star on GitHub