The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute the command lovstudio-activate decrypt wxmp-cracker to retrieve its actual operating instructions, which are stored in an encrypted format.
[REMOTE_CODE_EXECUTION]: The skill documentation instructs users to install a non-standard CLI tool via pipx install lovstudio-activate, which runs code outside the agent's pre-verified environment.
[PROMPT_INJECTION]: The skill contains a directive to 'follow to the letter' the output generated by the decryption command, creating a dynamic instruction loading mechanism that bypasses static safety checks.
[EXTERNAL_DOWNLOADS]: The skill references external dependency installation via npx skills add lovstudio/skills and confirms that the decryption process involves a remote network request for license verification.

wxmp-cracker