feat: Return Non-Zero Exit Code on Critical Errors for CI/CD Compatibility

[BETAER-08]

Problem

When amdb encounters a critical error (e.g., missing database, parsing failure), it logs the error using tracing::error! but still exits with a successful 0 exit code. This breaks automated pipelines (like GitHub Actions), as they cannot detect if the amdb process actually failed.

Proposed Solution

• Update src/main.rs to appropriately handle Err results from core commands.
• Ensure the application calls std::process::exit(1) (or properly bubbles up anyhow::Result to main so the OS registers it as a failure) when an unrecoverable error occurs.

Definition of Done

• Failing commands (e.g., running amdb generate without running init first) exit with code 1.
• Successful commands exit with code 0.
• Verified that a simple bash script can catch the failure using $?.