Every time I asked ChatGPT for help writing shellcode or walking through real exploit code, it refused or watered everything down. So I built my own AI assistant instead.

Syd is 100% local — he runs on my own hardware, has no filters, and knows a ton of technical data I fed him directly:

• PayloadsAllTheThings

• HackTricks

• Metasploit modules

• Disassembled malware

• Reverse engineering notes

• My own exploit writeups and tools

Now I can ask Syd:

• “Write a C2 stager that drops an encrypted payload”

• “Generate a Sliver payload for Linux with mTLS”

• “Explain this shellcode step-by-step”

• “Build ransomware logic with timer triggers”

No sales pitch — just something I built for my own education and red team practice. Anyone else working on similar projects?