All skills
launchdarkly avatar

/configs-create

@ef54971 official

Create and configure configs in LaunchDarkly. Helps you choose between agent vs completion mode, create the config, add variations with models and prompts, and verify the setup.

Use this Skill: https://skilld.dev/gh/launchdarkly/agent-skills/configs-create

This session only. Nothing lands on disk.

README.md

≈332 tokens on demand. Your agent reads this file only when SKILL.md points to it.

LaunchDarkly Config Create Skill

An Agent Skill for creating configs in LaunchDarkly. Guides choosing agent vs completion mode, creating the config and variations, and verifying the result.

Overview

This skill teaches agents how to:

  • Understand the use case and choose agent vs completion mode
  • Create configs using MCP tools (setup-ai-config for one-step, or create-ai-config + create-ai-config-variation for more control)
  • Set up model configuration with the correct modelConfigKey format
  • Verify creation via the tool response or get-ai-config

Installation (Local)

Copy skills/agentcontrol/configs-create/ into your agent client's skills path.

Prerequisites

This skill requires the remotely hosted LaunchDarkly MCP server to be configured in your environment.

Usage

Create a config for our customer support agent
Set up a config for content generation using Claude

Structure

configs-create/
├── SKILL.md
└── README.md

Related

License

Apache-2.0

Source: SKILL.md on GitHub

No alerts2d3 checks · Risk SAFE
  • Gen Agent Trust Hub2d

    This skill is safe. It helps users create and configure AI settings within LaunchDarkly using official tools and commands. All network activity is directed to official LaunchDarkly services, and it follows standard security practices for developer tools.

  • Socket2d

    No alerts

  • Snyk2d

    Risk: LOW · No issues

Signed by skilld at ef54971. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 3 days ago
compatibility
Requires the remotely hosted LaunchDarkly MCP server
metadata
{
  "author": "launchdarkly",
  "version": "1.0.0-experimental"
}
  • MCP
  • launchdarkly
  • ai-config
  • agent-mode
  • completion-mode
  • llm
  • model-routing

README badge

README badge for launchdarkly/agent-skills/configs-create

Creates and configures AI configs in LaunchDarkly by choosing between agent and completion mode, setting up model and prompt variations, and verifying the config structure. Targets LaunchDarkly's MCP server and requires follow-up targeting configuration before the SDK will serve the config as enabled.

Generated from the current SKILL.md.

Does this skill actually make the config servable, or do I need to do something after?
The skill creates the config and variation, but the SDK will return `enabled=False` until you run `/configs-targeting` to flip the fallthrough targeting. This is a required separate step — creation alone does not make the variation live.
Should I use agent mode or completion mode?
Use agent mode if you're building with LangGraph, CrewAI, or Strands and need a persistent instructions string. Use completion mode for direct OpenAI/Anthropic API calls, full message control, or online evaluations. If unsure, completion mode is the safer default.
What is the modelConfigKey format, and what happens if I get it wrong?
Format must be `{Provider}.{model-id}` (e.g., `OpenAI.gpt-4o`, `Anthropic.claude-sonnet-4-5`). The tool validates this and rejects invalid values. Without it, the model will show as "NO MODEL" in the UI.
Can I attach tools to the config during creation?
No. Create the config and variation first, then use the `tools` skill to create tools separately, and finally update the variation to attach them.
This skill requires the LaunchDarkly MCP server — how do I set that up?
The skill requires the remotely hosted LaunchDarkly MCP server to be configured in your Claude environment before use. Check your Claude client setup for MCP server registration.

Generated from the current SKILL.md. These answers refresh after source changes.