The test runs and generates true PASS and true FAIL results as appropriate. It is permissible for a test to FAIL even if the expected result is PASS if the software under test has a known defect that has been reported. The applicable bug number should be referenced in the error message so that it is easy to research the failure.
Watch for bogus success values
The test verifies PASS and FAIL results (versus returning the success or failure from a particular shell command... many shell commands return success because they successfully ran, not that they returned expected data. This usually requires user verification)
Security review
A cursory review of the code should be performed to make sure it does not contain obviously malicious or suspicious routines which appear more focused on damaging or casing the testing infrastructure versus performing a valid test.