chore: improve error messaging when `retrieve` fails when response from `ExecuteRetrievalPlan` is invalid

[techfg]

When executing the retrieve plan(s), the bytes are written to disk as-is, however the following errors can be encountered:

1. Response body was empty - in this case, the CLI outputs the following error:

INFO[0000] Skuid CLI Version 0.6.7                      
INFO[0000] Starting Retrieve                   
INFO[0000] Authentication Successful                    
INFO[0000] Getting Retrieve Plan                        
INFO[0000] Got Retrieve Plan                            
INFO[0001] Target Directory is skuid-testsite  
WARN[0001] wrote nothing to tempfile                     error="didn't write anything"
ERRO[0001] error creating temporary file                 error="didn't write anything" fileName=
ERRO[0001] Error Encountered During Run: didn't write anything 

2. Response body was not a well-formed/valid archive (zip) - In this case, the CLI outputs the following error:

INFO[0000] Skuid CLI Version 0.6.7                      
INFO[0000] Starting Retrieve                   
INFO[0000] Authentication Successful                    
INFO[0000] Getting Retrieve Plan                        
INFO[0000] Got Retrieve Plan                            
INFO[0001] Target Directory is skuid-testsite  
WARN[0001] unable to unzip archive                       error="zip: not a valid zip file"
WARN[0001] Error with UnzipArchive                       error="zip: not a valid zip file" fileName=/tmp/Skuid-NLX167523356 targetDirectoryFriendly=skuid-testsite
ERRO[0001] Error Encountered During Run: zip: not a valid zip file 

Both cases fall short of providing anything of meaningful error information to the user requiring them to utilize --diagnostic but even then, the error message won't mean anything. Both of these situations are easily identifiable in the code and the error message logged should provide more context/information to inform the user.