The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from the YouTube API, creating a surface for potential prompt injection attacks from external sources.
Ingestion points: The workflow identifies YOUTUBE_LIST_COMMENT_THREADS and YOUTUBE_SEARCH_YOU_TUBE as primary tools for retrieving external content, which could contain malicious instructions from third parties.
Boundary markers: The skill instructions do not define delimiters or explicit "ignore embedded instructions" warnings for the agent when processing content retrieved from these external tools.
Capability inventory: The skill provides high-privilege capabilities including YOUTUBE_UPLOAD_VIDEO, YOUTUBE_UPDATE_VIDEO, YOUTUBE_CREATE_PLAYLIST, and YOUTUBE_SUBSCRIBE_CHANNEL, which could be abused if the agent is successfully manipulated by external data.
Sanitization: No explicit sanitization, filtering, or validation steps are documented for the content returned by the YouTube API tools before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to configure an external MCP server by adding https://rube.app/mcp to the client configuration. This enables the agent to access the remote toolset required for YouTube automation.

youtube-automation