The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill connects to a remote Model Context Protocol (MCP) server at https://mcp.artlist.io/mcp. This is a reference to an official service provider's endpoint for media generation tools.
[COMMAND_EXECUTION]: The documentation in references/connect.md provides a standard configuration command (claude mcp add) for users to register the Artlist MCP server within their local agent environment. This is an administrative setup step for the tool rather than arbitrary command execution.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided briefs and fetches dynamic model rosters from a remote server, which constitutes a standard attack surface for indirect prompt injection.
Ingestion points: User briefs describing requested media (subject, mood, format) and the live model catalog fetched from https://mcp.artlist.io/mcp at the start of each session.
Boundary markers: The instructions do not define specific delimiters for input briefs but emphasize that the server maps intent to model parameters.
Capability inventory: Remote media generation (image and video) and consumption of paid account credits.
Sanitization: No specific sanitization methods for the user brief are mentioned, as the data is passed to a specialized generation server. The skill mitigates risks through explicit human review requirements for bulk generation and spend gates.

artlist-mcp