The problem is even worse than that. OpenAI and Anthropic have your source code and superior knowledge of its vulnerabilities. All you can do is hope that they won't one day use it against you.