All skills
launchdarkly avatar

/configs-variations

@3f246b4 official

Experiment with configs by creating and managing variations. Helps you test different models, prompts, and parameters to find what works best through systematic experimentation.

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

This session only. Nothing lands on disk.

README.md

≈304 tokens on demand. Your agent reads this file only when SKILL.md points to it.

LaunchDarkly Config Variations Skill

An Agent Skill for creating and managing config variations to experiment with different models, prompts, and parameters.

Overview

This skill teaches agents how to:

  • Design experiments (model comparison, prompt optimization, parameter tuning)
  • Create variations using clone-ai-config-variation (recommended) or create-ai-config-variation
  • Verify variations exist with correct configuration via get-ai-config

Installation (Local)

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

Prerequisites

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

Usage

Add a GPT-4o-mini variation to test cost savings
Create variations to compare Claude vs GPT-4 for our agent

Structure

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

Related

License

Apache-2.0

Source: SKILL.md on GitHub

No alerts2mo3 checks · Risk SAFE
  • Gen Agent Trust Hub2mo

    The skill provides instructions for managing LaunchDarkly AI configurations and experiments using a set of dedicated MCP tools. It guides the agent in creating, cloning, and verifying configuration variations for model comparison and optimization. No security risks, malicious patterns, or unsafe practices were detected.

  • Socket2mo

    No alerts

  • Snyk2mo

    Risk: LOW · No issues

Signed by skilld at 3f246b4. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated 2 months ago
compatibility
Requires the remotely hosted LaunchDarkly MCP server
metadata
{
  "author": "launchdarkly",
  "version": "1.0.0-experimental"
}
  • MCP
  • launchdarkly
  • experimentation
  • config-management
  • model-testing
  • prompt-optimization
  • ai-parameters
  • variations

README badge

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

Creates and manages AI config variations within LaunchDarkly to test different models, prompts, and parameters one variable at a time. Use this to systematically experiment with cost-reduction, quality-improvement, or latency-optimization through the LaunchDarkly MCP server's clone and create tools, keeping a baseline variation intact for comparison and rollback.

Generated from the current SKILL.md.

What MCP server does this skill require?
The remotely hosted LaunchDarkly MCP server. It must be configured in your environment before using this skill.
Can I test multiple variables at once in a single variation?
No. The skill enforces testing one variable per variation — change either model, prompt, or parameters, not combinations. This isolates what caused any performance difference.
Should I modify the baseline variation or create a new one?
Always create a new variation alongside the baseline. Never modify or delete the baseline — keeping it enables comparison and safe rollback if needed.
What fields should I pass when cloning a variation?
Pass only the fields you are testing: sourceVariationKey, key, name, and the override field (modelConfigKey, instructions, or parameters). Omit all other fields so they inherit from the source automatically.
How do I verify a variation was created correctly?
If you used clone-ai-config-variation, the response includes both source and created variations for immediate comparison. Otherwise, use get-ai-config to confirm the new variation matches your intent.

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