The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill handles the installation of the meitu-kaipai-cli package via npm install to ensure the required vendor-provided tools are present. This operation uses a specific version (0.1.11) and follows standard global installation patterns.
[COMMAND_EXECUTION]: The skill runs a setup script (ensure-cli.mjs) and calls the kaipai command to execute video processing tasks. These commands are executed as subprocesses with arguments passed as arrays, reducing the risk of shell-based command injection.
[DATA_EXFILTRATION]: The skill downloads media from user-provided URLs and uploads them to the vendor's CDN. These actions are the primary mechanism for the tool's intended functionality (video watermark removal) and are directed to the vendor's infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill processes external media files and URLs, which serves as an entry point for untrusted data.
Ingestion points: Media files and URLs provided in user requests or context (SKILL.md).
Boundary markers: None explicitly defined; the skill delegates processing directly to the vendor's CLI.
Capability inventory: Invokes subprocesses via meitu-kaipai-cli for processing and ensure-cli.mjs for environment setup (scripts/ensure-cli.mjs).
Sanitization: CLI calls use structured argument arrays in spawnSync rather than shell strings (scripts/ensure-cli.mjs).

kaipai-video-remove-watermark