The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the claude mcp add command-line interface to execute shell commands for server configuration. It also involves running npx and docker to initiate server processes.
[EXTERNAL_DOWNLOADS]: Fetches and executes MCP server implementations from official and well-known sources, including the Model Context Protocol organization, Upstash, Exa, and GitHub's container registry.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it takes untrusted user input—such as API keys, directory paths, and custom server commands—and interpolates them directly into shell command templates without explicit sanitization or shell escaping.
Ingestion points: User input gathered via AskUserQuestion for API keys, tokens, file paths, and custom command strings.
Boundary markers: Absent; user inputs are inserted directly into bash blocks.
Capability inventory: Shell command execution via claude mcp add, npx, and docker across multiple configuration steps.
Sanitization: None provided in the instructions; relies on the agent's default handling of interpolated variables.

mcp-setup