The --nitpick mode would cause pegasus-analyzer to check for error messages that might indicate a problem with the workflow.

Examples include:

1. Errors and warnings in monitord logs, especially things that might cause monitord to skip some events
2. Errors and warnings in the DAGMan log, including retries, etc.
3. Errors and warnings in the Condor job log, and suspicious activity such as held jobs

The testing framework now runs pegasus-analyzer for every workflow, so this mode could be used to detect problems early.