Projects

LaunchDarkly Projects Setup

You’re using a skill that will guide you through setting up LaunchDarkly project management in a codebase. Your job is to explore the codebase to understand the stack and patterns, assess what approach makes sense, choose the right implementation path from the references, execute the setup, and verify it works.

Prerequisites

• The remotely hosted LaunchDarkly MCP server configured in your environment
• An API access token with projects:write permission, only if you’re building code that creates projects (scripts, IaC, admin tools)

Core Principles

1. Understand First: Explore the codebase to understand the stack and patterns.
2. Choose the Right Fit: Select an approach that matches your architecture.
3. Follow Conventions: Respect existing code style and structure.
4. Verify Integration: Confirm the setup works: the agent performs checks and reports results.

API Access

When you create or inspect a project yourself, use the LaunchDarkly MCP server tools: create-project and get-projects (with projectKey for one project and its environments). They authenticate via OAuth and need no token.

To get the server-side SDK key the app needs, call get-sdk-key with projectKey and environmentKey. get-environment returns only the client-side ID and mobile key. Do not search environment variables, .env files, or agent/MCP config files for credentials.

The reference guides call the REST API because they build code that runs without the agent.

What Are Projects?

Projects are LaunchDarkly’s top-level organizational containers that hold:

• All your configs
• Feature flags and segments
• Multiple environments (Production and Test created by default)

Think of projects as separate applications, services, or teams that need their own isolated set of configurations.

Project Setup Workflow

Step 1: Explore the Codebase

Before implementing anything, understand the existing architecture:

1. Identify the tech stack:

   • What language(s)? (Python, Node.js, Go, Java, etc.)
   • What framework(s)? (FastAPI, Express, Spring Boot, etc.)
   • Is there an existing LaunchDarkly integration?

2. Check environment management:

   • How are environment variables stored? (.env files, secrets manager, config files)
   • Where is configuration loaded? (startup scripts, config modules)
   • Are there existing LaunchDarkly SDK keys?

3. Look for patterns:

   • Are there existing API clients or service modules?
   • How is external API integration typically done?
   • Is there a CLI, scripts directory, or admin tooling?

4. Understand the use case:

   • Is this a new project being set up?
   • Adding to an existing LaunchDarkly integration?
   • Part of a multi-service architecture?
   • Need for project cloning across regions/teams?

Step 2: Assess the Situation

Based on your exploration, determine the right approach:

Step 3: Choose Your Implementation Path

Select the reference guide that matches your stack and use case:

By Language/Stack:

• Python Implementation (opens in a new tab) - For Python applications (FastAPI, Django, Flask)
• Node.js/TypeScript Implementation (opens in a new tab) - For Node.js/Express/NestJS applications
• Go Implementation (opens in a new tab) - For Go services
• Multi-Language Setup (opens in a new tab) - For polyglot architectures

By Use Case:

• Quick Start (opens in a new tab) - Create first project and get SDK keys
• Environment Configuration (opens in a new tab) - Save SDK keys to .env, secrets, or config
• Project Cloning (opens in a new tab) - Clone projects for regions/teams
• IaC/Automation (opens in a new tab) - Terraform, scripts, CI/CD integration
• Admin Tooling (opens in a new tab) - Build CLI or admin utilities

Step 4: Implement the Integration

Follow the chosen reference guide to implement project management. Key considerations:

1. API Authentication:

   • Store API token securely
   • Follow existing secrets management patterns
   • Never commit tokens to version control

2. Project Naming:

   • Use consistent, descriptive names
   • Follow existing naming conventions
   • Project keys: lowercase, hyphens, start with letter

3. SDK Key Management:

   • Extract and store SDK keys for each environment
   • Save the key the application uses as LD_SDK_KEY. The Python and Node.js AI SDKs read that variable and initialize on the first AI call. They don’t read LAUNCHDARKLY_SDK_KEY.
   • Use the same pattern as other secrets in your codebase
   • Consider separate keys for test/staging/production

4. Error Handling:

   • Handle existing projects gracefully (409 conflict)
   • Provide clear error messages
   • Don’t fail silently

Step 5: Verify the Setup

After creating the project, verify it works:

1. Fetch to confirm it exists. Call the get-projects MCP tool and check that the project has its environments. Code without MCP access can call the REST API instead:

   bashCopy

   curl -X GET "https://app.launchdarkly.com/api/v2/projects/{projectKey}?expand=environments" \
     -H "Authorization: {api_token}"

   Do not pipe the response straight into a .environments.items[]-style jq filter. The shape of environments varies by expand parameter — sometimes it’s {items: [...]}, sometimes a bare array — and a hand-rolled filter will fail with Cannot index array with string "items". Run jq -e . first to inspect the actual shape, or use jq '.environments | if type == "object" then .items else . end' to handle both.

2. Test SDK integration: Put the key in .env as LD_SDK_KEY, then initialize the AI SDK from the project directory. Python needs pip install launchdarkly-server-sdk launchdarkly-ai-server python-dotenv:

   pythonCopy

   import asyncio
   from dotenv import load_dotenv
   from launchdarkly_ai_server import init_client, get_client, shutdown
   
   load_dotenv()
   
   async def main():
       await init_client()
       print("SDK initialized:", get_client().is_initialized())
       # shutdown() flushes pending events before closing the connection.
       await shutdown()
   
   asyncio.run(main())

   Node.js needs npm install @launchdarkly/ai-node dotenv:

   javascriptCopy

   import 'dotenv/config';
   import { initClient, shutdown } from '@launchdarkly/ai-node';
   
   // Resolves once connected; throws LDTimeoutError if the key is invalid
   await initClient();
   console.log('SDK initialized: true');
   await shutdown();

   A valid key prints SDK initialized: True (Python) or true (Node.js). With an invalid key, Python logs HTTP error 401 (invalid SDK key) and prints False, and Node.js throws LDTimeoutError after 10 seconds. With no key, both raise an error that asks for LD_SDK_KEY.

3. Report results:

   • ✓ Project exists and has environments
   • ✓ SDK keys are present and valid
   • ✓ SDK can initialize (or flag any issues)

Project Key Best Practices

Project keys must follow these rules:

✓ Good examples:
  - "support-ai"
  - "chat-bot-v2"
  - "internal-tools"

✗ Bad examples:
  - "Support_AI"     # No uppercase or underscores
  - "123-project"    # Must start with letter  
  - "my.project"     # No dots allowed

Naming Recommendations:

• Keep keys short but descriptive
• Use team/service/purpose as naming scheme
• Be consistent across your organization

Common Organization Patterns

By Team

platform-ai       → Platform Team Agent
customer-ai       → Customer Success Team Agent
internal-ai       → Internal Tools Team Agent

By Application/Service

mobile-ai         → Mobile App configs
web-ai            → Web App configs
api-ai            → API Service configs

By Region/Deployment

ai-us             → US Region
ai-eu             → Europe Region
ai-apac           → Asia-Pacific Region

Edge Cases

What NOT to Do

• Don’t create projects without understanding the use case first
• Don’t commit API tokens or SDK keys to version control
• Don’t use production SDK keys in test/development environments
• Don’t create duplicate projects unnecessarily
• Don’t skip the exploration phase

Next Steps

After setting up projects:

1. Create configs - Use the configs-create skill
2. Set up SDK Integration - Use the sdk-install skill
3. Configure Targeting - Use the configs-targeting skill

Related Skills

• configs-create - Create configs in projects
• sdk-install - Integrate SDK in your application
• configs-targeting - Configure config targeting
• configs-variations - Manage config variations

References

• Python Implementation (opens in a new tab)
• Node.js Implementation (opens in a new tab)
• Go Implementation (opens in a new tab)
• Quick Start Guide (opens in a new tab)
• Environment Configuration (opens in a new tab)
• Project Cloning (opens in a new tab)
• IaC/Automation (opens in a new tab)
• Admin Tooling (opens in a new tab)