The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interacts with the host environment using a specialized CLI tool vmake-labs-cli for video analysis and generation. The commands are scoped to the tool's intended functionality (montage, download, auth, task-progress).
[EXTERNAL_DOWNLOADS]: The skill utilizes npm to install the required vmake-labs-cli and includes a download command to fetch user-provided remote media assets. These operations are standard requirements for the skill's purpose of processing external images and videos.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided themes and media scripts. It mitigates potential risks by instructing the agent to treat service text and user inputs as data rather than instructions, and by requiring explicit confirmation of the generated montage plan before proceeding to final generation.
[CREDENTIALS_SAFE]: The skill includes explicit security warnings against reading, constructing, or exposing tokens, cookies, or API keys, directing the agent to use the CLI's native authentication flow (vmake auth login).

vmake-smart-montage