The Agent Skills Directory

[SAFE]: No malicious patterns such as prompt injection, data exfiltration, or obfuscation were detected within the skill content.
[EXTERNAL_DOWNLOADS]: The skill recommends installing the @membranehq/cli package via npm and uses npx to run commands from the same package. These are official tools provided by the vendor ('membranedev') for interacting with their infrastructure and are considered safe within this context.
[SAFE]: The authentication process described utilizes the membrane login and membrane connect flows, which avoid exposing sensitive credentials directly to the agent or local environment by handling them server-side.

reviewflowz