The Agent Skills Directory

[SAFE]: The skill instructs the agent to avoid hardcoding credentials and instead use environment variables (e.g., LAUNCHDARKLY_SDK_KEY, LAUNCHDARKLY_CLIENT_SIDE_ID), which is a security best practice for secret management.
[SAFE]: The integration plan explicitly includes steps to ensure that sensitive environment files (like .env) are created only with user consent and are properly listed in .gitignore to prevent accidental exposure in version control.
[SAFE]: The skill implements a 'BLOCKING' mechanism that halts execution until the user provides explicit approval of the proposed integration plan, ensuring human-in-the-loop oversight.
[SAFE]: References to external documentation and package names are confined to official LaunchDarkly resources, consistent with the vendor's own tools and services.

plan