Harness FME FAQs

Name: Harness FME
Author: kud

Question 1

What is Harness FME and its primary use?

Accepted Answer

Harness FME is an MCP (Model Context Protocol) server designed to enable AI assistants to manage Harness FME (Split.io) feature flags. It allows AI assistants to read, toggle, kill, and restore feature flags programmatically across various workspaces and environments.

Question 2

Are there safety measures for critical operations like 'killing' a feature flag?

Accepted Answer

Absolutely. Harness FME includes a 'Safety Guard' for kill operations. To prevent accidental flag kills, it requires an explicit confirmation parameter (e.g., `confirm: true`) to be passed before disabling a flag.

Question 3

Which AI assistants and IDEs can I integrate Harness FME with?

Accepted Answer

Harness FME offers native integration via the MCP protocol with a wide range of AI assistants and IDEs, including Claude Desktop, Claude Code CLI, Cursor, VSCode (with compatible extensions), Windsurf, and all JetBrains IDEs.

Question 4

Can I manage feature flags across different environments and workspaces?

Accepted Answer

Yes, Harness FME provides multi-workspace support, allowing you to interact with any number of Harness FME workspaces. It also supports environment-aware operations, enabling you to target feature flags in specific environments like staging or production.

Question 5

What specific feature flag actions can my AI assistant perform?

Accepted Answer

Your AI assistant can perform a variety of actions including listing all feature flags, inspecting a specific flag, getting its definition and targeting rules, killing (disabling) a flag, and restoring (re-enabling) a killed feature flag.

Harness FME

Harness FME

主な機能

ユースケース

主な機能

ユースケース