The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the official membrane CLI tool to execute actions, manage connections, and handle NS1 resources.
[EXTERNAL_DOWNLOADS]: The instructions include downloading the @membranehq/cli package from the public npm registry, which is the official tool provided by the vendor.
[PROMPT_INJECTION]: A surface for indirect prompt injection exists in SKILL.md. Untrusted data enters the agent context via the intent parameter in the action list command and the input JSON in the action run command. While these are necessary for the skill's functionality, no explicit boundary markers or sanitization procedures are defined to mitigate malicious instructions embedded in these user-controlled inputs.

ns1