Nuclei -o Flag Overwrites Previous Results

[skalvino]

When running Nuclei with the -o flag, the output file is overwritten with the latest results instead of appending to the existing file.
This causes a loss of previous scan results.

Steps to Reproduce:

Run a Nuclei scan with the -o flag specifying the output file.

nuclei -u http://example.com -o results.txt

Run another Nuclei scan with the -o flag specifying the same output file.

nuclei -u http://anotherexample.com -o results.txt

Observe that the results.txt file contains only the results from the latest scan, overwriting the previous results.

[GeorginaReeder]

Hi @skalvino !

Thanks for this - could you please go ahead and create a bug report? We can take a closer look from there.

[skalvino]

Thank you for your response; as Ehsandeep mentioned, maybe we can use

nuclei -u http://anotherexample.com | tee -a results.txt to resolve the issue with using -o

However, people who use nuclei with the default -o may not realize that they are losing data results if they use it in pipe mode, for example.

cat urls.txt | nuclei -t template.yaml --output.txt

Actually, it's not a bug, but rather an idea to make the nuclei work more efficiently

[ehsandeep]

@skalvino This is the default and expected behavior. Overwriting existing files can cause multiple issues, as it may overwrite files unrelated to Nuclei results. While appending results is possible in various ways, keeping overwrite as the default is not the best option.

Here’s an example of how you can append results to an existing file:

nuclei -u http://anotherexample.com | tee -a results.txt

[skalvino]

If you use nuclei with pipe mode,

cat urls.txt | nuclei -t template.yaml -o output.txt

In this way, it will overwrite the result with the last scan, and you will lose the data of the scan if you are working on a large scope or a long list of URLs.

That's why I'm sharing the idea of having it append results to an existing file.

[ehsandeep]

Appreciate you sharing the idea, but there are more problems than solutions in making this the default behavior. Instead, you can randomize the output file name like this:

cat urls.txt | nuclei -o "output_$(date +%s)_$RANDOM.txt"