Skip to content

Security: asgard-ai-platform/mcp-uof

Security

SECURITY.md

Security Policy

Supported Versions

This repository is currently pre-1.0. Security fixes are applied to the latest main branch.

Reporting a Vulnerability

Do not open a public issue for secrets, credentials, tokens, authorization bypasses, or data exposure reports.

Report privately to the repository maintainers through your normal security contact channel. Include:

  • A clear description of the issue.
  • A minimal reproduction if possible.
  • Affected files, commands, or configuration.
  • Whether any credential, token, hostname, user data, or customer data was exposed.

Credential Handling

  • Never commit .env, UOF accounts or passwords, RSA key pairs, generated tokens, generated reports, or local credential files.
  • The MCP server operates with the plaintext UOF_ACCOUNT / UOF_PASSWORD from .env; the code RSA-encrypts them before sending. Keep .env out of version control.
  • Use .env.example for placeholders only.
  • Treat manual and end-to-end test outputs as sensitive unless reviewed and sanitized.

Deployment Notes

  • UOF_VERIFY_SSL=false is only intended for test environments with incomplete or self-signed certificates.
  • Use strict SSL verification in production.
  • The SSE demo token mechanism (MCP_SSE_TEST_TOKEN) is for local testing. Replace it with real token verification before deploying an exposed service.

There aren't any published security advisories