asmitaj08 / FuzzingBusyBox_LLM

MIT License
6 stars 1 forks source link

what is the /path/to/corpus? #1

Open Dejavu610 opened 2 months ago

Dejavu610 commented 2 months ago

Dear author,

Could you please provide an example of the corpus?

asmitaj08 commented 2 months ago

It is the path of dir that has busybox elfs. You can download some sample firmware from camera/router vendor website like : DLink, Netgear, Tenda ,etc. You would have to extract the binary to fetch BusyBox ELF (common tool to extract- binwalk)

Dejavu610 commented 2 months ago

It is the path of dir that has busybox elfs. You can download some sample firmware from camera/router vendor website like : DLink, Netgear, Tenda ,etc. You would have to extract the binary to fetch BusyBox ELF (common tool to extract- binwalk)

Thank you for your prompt reply. I see the help information " --input", dest='target_binary', help =' Enter the path of the target binary that you wnat to fuzz', required=True," in the in the fuzz_one_target_main.py. It seems that --input is the path to of busybox elfs and corpus is the path of seed. I wonder what the format of seed should be? Is it possible to provide an example of a gpt generated seed?

Thank you very much.

asmitaj08 commented 2 months ago

Yes, my bad. Replied in rush, and misread it. It's for the collection of input seeds. It could be either random or LLM generated. Here are samples for some seeds for Busybox awk utility generated by LLM . The attached example shows some samples

Screenshot 2024-07-12 at 3 43 16 AM

. You can write them in different files and store them as seed in the input_corpus dir.