mcp-server-manager
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill contains numerous examples of shell commands (e.g.,
claude mcp list,claude mcp add). These are provided as instructional documentation for the user to manually execute in their terminal and do not represent hidden or automated execution by the agent. - [EXTERNAL_DOWNLOADS] (SAFE): Documentation includes examples of using
npx -yto install MCP server packages. While this involves downloading code, it is the standard and intended method for the described tool's functionality. - [CREDENTIALS_UNSAFE] (SAFE): The skill demonstrates how to use headers and environment variables for authentication. It correctly uses non-sensitive placeholders such as
YOUR_TOKEN,YOUR_KEY, andsecret123in its examples, which is consistent with security best practices for documentation. - [DATA_EXPOSURE] (SAFE): The skill mentions configuration file paths like
~/.claude/settings.jsonand~/.mcp-auth/. These references are educational, explaining where the tool stores its own configuration, rather than attempting to exfiltrate sensitive user data.
Audit Metadata