The Agent Skills Directory

[PROMPT_INJECTION]: The skill's core operational instructions are hidden in an AES-256-GCM encrypted payload (SKILL.md.enc), which prevents security review of the agent's instructions. The skill explicitly directs the agent to adopt and follow instructions fetched dynamically from the output of an external command at runtime.\n- [REMOTE_CODE_EXECUTION]: The skill instructs the user to execute uvx lovstudio-skill-helper, a command that downloads and runs a Python package directly from a remote registry. This bypasses local safety checks and allows for arbitrary code execution.\n- [COMMAND_EXECUTION]: The skill provides the agent with specific shell commands to guide the user through software installation, decryption, and license activation processes.\n- [EXTERNAL_DOWNLOADS]: The skill facilitates the download of encrypted content bundles and third-party dependencies from vendor-controlled servers and external package registries (NPM and PyPI).\n- [DATA_EXFILTRATION]: The activation process directs users to provide license keys and log in via the lovstudio.ai domain, which involves the transmission of potentially sensitive user-purchased credentials to an external third-party server.

lovstudio-event-curator