The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill executes npx -y mcp-video-analyzer@latest, which downloads and runs a package from the npm registry at runtime. This package is maintained by the skill's author and is the primary tool for video analysis in environments without a dedicated MCP server.
[COMMAND_EXECUTION]: The skill uses the Bash tool to run shell commands for video processing, package execution, and interacting with the mcp-video-analyzer CLI.
[INDIRECT_PROMPT_INJECTION]: The skill extracts and processes data from external videos (transcripts, OCR, metadata), which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Analysis output from mcp-video-analyzer containing timestamped transcripts, on-screen OCR text, and video metadata in SKILL.md.
Boundary markers: Absent. The instructions do not define specific delimiters to separate the extracted video content from the agent's instructions.
Capability inventory: The agent possesses Bash execution privileges, the ability to Read local files (for frame inspection), and access to the mcp__video-analyzer toolset.
Sanitization: The skill performs initial URL validation to block access to private/LAN addresses and metadata endpoints, though this can be overridden by user configuration.
[EXTERNAL_DOWNLOADS]: At runtime, the skill downloads dependencies from the npm registry and uses yt-dlp to fetch media from external platforms like YouTube, Instagram, and TikTok.

video