The Agent Skills Directory

[CREDENTIALS_SAFE]: The skill implements strong security practices for secret management. It instructs the agent to create an API key using the vendor CLI, export it to an environment variable ($WARP_API_KEY), and use that variable in all subsequent commands and configuration files (like mcp.json or config.toml). It explicitly warns the agent never to paste raw keys into files or print them to logs.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute standard configuration commands for external tools like claude, codex, and curl. These commands are used to register a Model Context Protocol (MCP) server.
[EXTERNAL_DOWNLOADS]: The skill references the oz CLI and an MCP server URL. Both resources originate from the vendor's official domains (warp.dev). The skill correctly uses a placeholder for the installation command of the oz CLI, preventing the agent from guessing or executing an unverified installation script.
[SAFE]: The skill's operations are scoped to the vendor's infrastructure and follow standard integration patterns for MCP-capable coding agents.

factory-setup