The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl to interact with the LiteLLM proxy's /v1/mcp/server endpoints. These commands are necessary for its stated purpose of listing and deleting server registrations.
[CREDENTIALS_UNSAFE]: The skill references environment variables LITELLM_BASE_URL and LITELLM_API_KEY. This is a recommended security practice to avoid hardcoding secrets and to ensure that sensitive information remains in the user's controlled environment.
[EXTERNAL_DOWNLOADS]: The skill communicates with an external LiteLLM proxy. As this interaction targets the user's own infrastructure or resources associated with the trusted vendor BerriAI, and no sensitive files are accessed or exfiltrated, this is considered safe behavior.
[INDIRECT_PROMPT_INJECTION]: While the skill processes data from an external API (listing servers), this administrative surface is standard for management tools. Given the trusted nature of the vendor and the specific use case, it does not present a significant risk factor.

delete-mcp