trufflesecurity/trufflehog ? reverse-engineered prompt

Build me a security tool for finding leaked credentials. I'm worried about developers accidentally committing API keys, database passwords, and other secrets directly into our codebases.

The tool should be able to scan entire GitHub organizations, specific repositories, and local file directories. It needs to be more than a simple text search; it should be able to identify what kind of secret it finds—like telling me if it's an AWS key versus a Stripe key, for example.

The most important feature is validation. After finding a potential secret, I want the tool to automatically check if the credential is still live and active. This will help us prioritize real threats instead of chasing down old, revoked keys. I imagine this as a command-line utility that I can run on demand, giving me a clear report of any active, leaked secrets and where to find them.