Skip to content

Security: ellmos-ai/ellmos-controlcenter-mcp

Security

SECURITY.md

Security

Alpha Status

ellmos-controlcenter-mcp is currently an alpha release. It can inspect profiles, prepare generated MCP configs, show local MCP servers, and audit first policy hints. It is not yet a security boundary.

Current Safety Model

  • Dashboard write actions require explicit confirmation.
  • Existing profile files and generated MCP config files are backed up before overwrite.
  • Environment values are not printed by the profile audit.
  • The dashboard binds to 127.0.0.1 by default.

Not Yet Implemented

  • Tool-level policy enforcement
  • MCP gateway/proxy enforcement
  • User authentication for the dashboard
  • Remote multi-user access control
  • Tamper-proof audit logs

Recommendations

  • Run the dashboard only on trusted local machines.
  • Keep Claude profiles small and task-specific.
  • Do not expose the dashboard host publicly.
  • Treat npx-based MCP servers as less reproducible than pinned local paths.
  • Review generated MCP configs before launching a client with them.

Reporting

For private preview issues, use the private GitHub repository:

https://github.com/ellmos-ai/ellmos-controlcenter-mcp/issues

There aren't any published security advisories