The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill provides explicit security guidance to prevent credential leakage. It instructs users to never hardcode API tokens and instead use environment variables like $LAUNCHDARKLY_ACCESS_TOKEN for shell commands and API requests.
[COMMAND_EXECUTION]: The skill utilizes curl and the ldcli command-line tool to perform legitimate operations such as creating and toggling feature flags via the official LaunchDarkly API (app.launchdarkly.com).
[EXTERNAL_DOWNLOADS]: Recommends the installation of auxiliary tools from LaunchDarkly's official GitHub repository (github.com/launchdarkly/ai-tooling) using npx, which is consistent with the skill's primary purpose and originates from a well-known service provider.

first-flag