Claude Code Tutorial: How to Connect Claude Code to a Task Board (2026)

You use Claude Code in your terminal. It writes code, debugs, refactors. But if you're running multiple sessions — one for backend, one for frontend, one for tests — you lose track of what each session is doing. Context bleeds between conversations. You manually paste requirements from one session into another.

This tutorial shows you how to connect Claude Code to a shared task board so you can assign tasks from a web dashboard, track progress across sessions, and even coordinate Claude Code with other AI agents (Cursor, OpenCode, OpenAI) in the same workspace.

What you'll learn:

• Why Claude Code needs task management
• How to connect Claude Code to Ivern Squads
• Running your first task
• Coordinating Claude Code with other agents
• Real workflows and prompts

Related: Getting Started with Ivern · Claude Code vs Cursor Comparison · OpenCode Tutorial · AI Coding Assistant Complete Guide · Get Your First AI Agent Result

Why Claude Code Needs Task Management

Claude Code is excellent at executing individual coding tasks. But real projects aren't individual tasks — they're coordinated sequences of work. Here's what happens without structure:

The unstructured workflow (what you're probably doing now):

1. Open Claude Code terminal
2. Type "build a REST API for user management"
3. Wait for output
4. Copy-paste requirements into a second Claude Code session for the frontend
5. Open a third session to write tests
6. Lose track of which session has the latest context
7. Manually reconcile changes

The structured workflow (what this tutorial enables):

1. Open Ivern dashboard in your browser
2. Create a task: "Build REST API for user management"
3. Assign it to your Claude Code agent
4. Claude Code picks it up and executes
5. Create a second task: "Build frontend forms matching the API"
6. Assign it to a second agent (could be Cursor or another Claude Code session)
7. Both agents work from the same task board — you see progress in one view

The difference: you manage tasks, not terminals.

How to Connect Claude Code to Ivern Squads

Prerequisites

• Claude Code installed on your machine (installation guide)
• Node.js 18+ (for the Ivern agent daemon)
• A free Ivern account

Step 1: Generate an Agent API Key (30 seconds)

1. Sign in to your Ivern dashboard
2. Go to Settings → Connections → Install
3. Click Generate Key
4. Name it (e.g., "My MacBook Pro")
5. Copy the key — it starts with ivern_ and won't be shown again

Step 2: Configure Claude Code (60 seconds)

Add the Ivern agent endpoint to your Claude Code settings. Edit .claude/settings.json:

{
  "externalAgents": {
    "ivern-squads": {
      "endpoint": "https://ivern.ai/api/agent-protocol",
      "authHeader": "Bearer $IVERN_API_KEY"
    }
  }
}

Set the environment variable:

export IVERN_API_KEY=ivern_YOUR_KEY_HERE

Verify the connection by starting Claude Code — tasks from Ivern will appear automatically when assigned to your squad.

Step 3: Create a Squad with Your Claude Code Agent (30 seconds)

1. In Ivern, click Create Squad
2. Name it "Dev Team"
3. Add an agent:
   • Name: Backend Dev
   • Role: Coder
   • Connection: Select your Claude Code connection
4. Click Create Squad

Your Claude Code session is now a squad member. It will pick up tasks assigned to this squad.

Running Your First Task

From your Ivern dashboard, open your "Dev Team" squad and click New Task. Try this prompt:

Refactor the user authentication module in src/auth/. Extract the JWT validation into a separate middleware function, add proper error handling for expired tokens, and write unit tests for the new middleware. The existing tests in src/auth/tests/ should all still pass.

Click Create Task. Your Claude Code agent picks it up and starts executing. You see the output stream in real time on the Ivern dashboard.

Cost: Approximately $0.02–$0.10 in Anthropic API costs. Ivern adds zero markup.

Coordinating Claude Code with Other Agents

This is where multi-agent coordination becomes powerful. You can mix Claude Code with other AI agents in the same squad.

Example: Full-Stack Dev Squad

Agent	Tool	Role	What It Does
Backend Dev	Claude Code	Coder	Writes API endpoints, database queries, tests
Frontend Dev	Cursor	Coder	Builds UI components, integrates with API
Reviewer	Claude (BYOK)	Reviewer	Reviews PRs, catches bugs, checks code quality

Create this squad in Ivern:

1. Create Squad → name it "Full-Stack Team"
2. Add Backend Dev → connect your Claude Code, role: Coder
3. Add Frontend Dev → connect Cursor, role: Coder
4. Add Reviewer → use BYOK with your Anthropic key, role: Reviewer

Now when you assign a task like "Build a user profile page with editable fields and avatar upload," the Backend Dev creates the API endpoints, the Frontend Dev builds the form, and the Reviewer checks everything before you see the output.

Example: Research + Code Squad

Agent	Tool	Role	What It Does
Researcher	Claude (BYOK)	Researcher	Gathers requirements, API docs, best practices
Coder	Claude Code	Coder	Implements based on the research
Writer	GPT-4o (BYOK)	Writer	Writes documentation for the code

Assign: "Research the Stripe payment intent API, implement a payment flow in our Express.js backend, and document the integration."

The Researcher gathers Stripe API docs and creates a spec. The Coder (Claude Code) implements the code. The Writer (GPT-4o) produces documentation. Each agent specializes in what it does best.

Real Workflows and Prompts

Workflow 1: Bug Triage

Squad: Single Claude Code agent

Prompt:

Review the following error logs from our production server. Identify the root cause of the 500 errors occurring in the /api/orders endpoint. Fix the issue in the relevant source file and add a regression test. Error logs: [paste logs]

Expected output: Root cause analysis + code fix + test file. Cost: $0.03–$0.08

Workflow 2: Feature Implementation Pipeline

Squad: Researcher (BYOK) + Coder (Claude Code) + Reviewer (BYOK)

Prompt:

Implement user rate limiting for our API. The requirements are: 100 requests per minute per user for free tier, 1000 per minute for paid tier. Use Redis for tracking. Include middleware, configuration, and tests. Follow our existing pattern in src/middleware/auth.ts.

Expected output: Complete feature with tests. Cost: $0.08–$0.20

Workflow 3: Code Migration

Squad: Claude Code agent

Prompt:

Migrate our database queries from raw SQL to Prisma ORM. Start with the models in src/models/, update all queries in src/routes/ to use Prisma client, and ensure the existing test suite passes. Do not change the API contract.

Expected output: Migrated codebase with passing tests. Cost: $0.10–$0.30 (depends on codebase size)

Cost Breakdown

Ivern is free. You pay only Anthropic API costs at direct provider pricing:

Task Type	Typical Input	Typical Output	Approx. Cost
Bug fix	2K tokens	1K tokens	$0.01–$0.03
Feature implementation	5K tokens	3K tokens	$0.05–$0.15
Code review	8K tokens	2K tokens	$0.03–$0.08
Full migration	20K tokens	10K tokens	$0.15–$0.40

Compare this to $20/month for a single ChatGPT Plus subscription. With BYOK, $5 in Anthropic credits covers 100+ coding tasks.

Frequently Asked Questions

Does this replace my terminal workflow?

No. Claude Code still runs in your terminal. Ivern adds a coordination layer on top. You can use Claude Code directly in the terminal for ad-hoc work, and use Ivern for structured tasks that need tracking, coordination, or collaboration with other agents.

What happens when Claude Code is disconnected?

Tasks remain queued. When Claude Code reconnects, it picks up pending tasks automatically. No work is lost.

Can I use this without the daemon?

Yes. Use the BYOK path instead — add your Anthropic API key directly in Ivern settings. Ivern calls Claude directly without needing a local agent. You lose the ability to use your local Claude Code session, but you get the same model through the API.

How is this different from using Claude Code directly?

Direct Claude Code: You type a prompt, it responds. One session, one task at a time.

Ivern + Claude Code: You assign tasks from a dashboard, multiple agents can work simultaneously, results are tracked and reviewable, and you can coordinate with other tools (Cursor, OpenAI) in the same workflow.

Is my code sent to Ivern?

No. When using the BYOA path (Claude Code daemon), your code stays on your machine. Ivern only receives the task description and the final output. The actual code execution happens locally in Claude Code.

Can I mix this with Cursor?

Yes. Connect both Cursor and Claude Code to the same Ivern squad. Claude Code handles backend tasks, Cursor handles frontend — all from the same task board. See the Claude Code vs Cursor comparison for when to use each.

Get Started

1. Sign up at ivern.ai/signup (free, no credit card)
2. Configure your Claude Code settings with the Ivern endpoint
3. Create a squad and assign your first task

Your Claude Code session becomes a managed, coordinated agent in 2 minutes.

For more workflows, check out 10 AI Agent Workflows You Can Set Up Today. For the full technical reference, see the Getting Started guide.

Connect Claude Code Now →