CC13: LLMs: Your Wireless Hacking Friend

See how large‑language models can speed up wireless penetration tests—from decoding pcaps to crafting payload templates—using nothing but well‑written prompts. LLMs aren’t limited to essays. In this talk Alissa Gilbert (“dnsprincess”) shows red‑teamers and hobby wardrivers how ChatGPT, Claude and other models cut analysis time during Wi‑Fi and BLE assessments. Learn prompt patterns that: - translate tcpdump / Wireshark logs into plain English, - identify four‑way handshakes and hidden SSIDs inside pcaps, - draft reverse‑shell scripts, router‑exploit PoCs and Metasploit commands, and - suggest targeted wordlists for WPA‑PSK cracking. You’ll also hear what AI can’t do—touch hardware, keep you legal, or replace fundamentals—and how to keep models on track with ethical, specific, context‑rich prompts. 00:00 Track‑host intro & session start 00:24 Speaker intro – Alissa “dnsprincess” Gilbert 01:02 Agenda & focus on example prompts 01:20 Wireless pentest basics (recon → attack) 01:49 What LLMs can do – log & pcap analysis, recon help 02:55 Limits of LLMs & prompt‑engineering must‑dos 03:15 Using ChatGPT safely – lab context, specificity, docs 03:53 Prompt demo 1 – crafting a Wi‑Fi reverse shell 04:24 Prompt demo 2 – pulling CVE data for a router model 04:46 Prompt demo 3 – encrypting payloads with AI assistance 05:02 Prompt demo 4 – extracting WPA/WPA2 handshakes 05:20 Prompt demo 5 – analyzing pcaps for hidden SSIDs 05:47 Workflow – feed AI output into aircrack‑ng & tools 06:07 Dictionaries, wordlists & CPY integration 08:53 Dragonblood (WPA3) prompt example 09:23 Tips for efficient, narrowly‑scoped prompts 12:49 Handling large pcaps & splitting data 13:32 Self‑hosted LLMs – current limits 14:55 Wrap‑up, links & audience Q&A #WirelessHacking #llm #CactusCon #WiFiPentest #promptengineering