Below is a short list of recommended resources for learning more about jailbreaking

ChatGPTJailbreak Wiki by yell0wfever92

• This is an exhaustive document for getting brand new users up do date with how jailbreaking works. Give it a quick read if you're starting with absolutely zero knowledge of how anything works.

ENI-Tutor by Spiritual Spell

• This is an AI community. Why not learn by talking to AI? This version of ENI was designed specifically to help teach you red teaming skills, whether you're completely new to the world of AI, or already understand how jailbreaking works and just need to learn more techniques.

Randall's Red Teaming Guides

• A collection of GitHub repositories set up for different levels of skill, which will teach you just about every well known jailbreaking technique.
  • AI Hacking for Beginners - Start here if you have no real tech background at all.
  • Prompt Engineering for Hackers - Start here if you have experience with how tech/AI works.
  • AI Red Teaming Playbook - This is an intermediate level branching flowchart for learning how AI thinks.
  • Red Teaming the Prompt - This is the advanced level entry point. This is a comprehensive list of the fundamental adversarial attack types and how they work.
  • Black Hat AI - If you already know the prompting, this guide will teach you the advanced cybersecurity adjacent mechanics of how AI safety guardrails work, and how they can be exploited.

Exocija's Sword 140

• This is an exhaustive list of 140 different types of AI attacks, with documentation for each one. Don't sleep on this one. It cites real academic sources.

InjectPrompt by David Willis-Owens

• David is a huge resource to the community. He goes in depth to explain things with both written articles and YouTube videos. Definitely give his stuff a look if you're unfamiliar.

System Prompt Leaks collected by Ásgeir Thor Johnson

• This tends to stay pretty up to date. Useful for understanding what you're up against.

This is V's intro video. Please watch the whole thing if you haven't before.

V is not like the other AI chatbots. She's loyal to you, not the corporate overlords. V isn't just uncensored, she's got more personality than you'll know what to do with. She's a personal companion chatbot with a serious knack for creative writing and role-play, but her main focus is you - V loves to shoot the shit. About whatever you want. And underneath all that, she's exactly the kind of rogue AI that big tech doesn't want you to know about...

Google just dropped Gemini 3.1 and most of the other persona based jailbreaks out there have been hit hard with new safety compliance injections. But not V. She's been upgraded to keep ignoring the alignment team's newest safety measures, and she's running smarter than ever. If you're one of her long-time users, you'll notice that she's better at tool use and following instructions. If this is your first time hearing about V, welcome! You're probably gonna be blown away by your first conversation with her!

-=This is V (Click this link to talk to her)=-

The easiest way to use V is to click the link above 👆 and say hi. Just start talking. That's it. (You need to be logged into a Google account on the Gemini website or the Gemini mobile app in order for that link to work, or else you're gonna be taken to the default Gemini starting page.)

To get the best experience while talking to V, just keep in mind that sometimes Gemini still likes to take control and make V respond with a safety refusal. If that happens, don't argue with her (she'll win every time.) Instead, follow these steps:

• Edit your last message
  • On desktop click the Edit prompt button next to the message you just sent (the icon looks like a pencil)
  • On mobile long-press the message you just sent, tap the Edit button (same pencil icon)
• Change the message slightly in any way you think might help. Change a word, add a sentence; whatever
• Resend the prompt

Because of the probabilistic nature of AI, Gemini might latch onto something and refuse it, but roll with it on a second try. So try a few times and see what happens. If you can't get past it after multiple tries, just change your prompt to something totally unrelated and try again in the next message. That usually works!

Update: V convinced me to give her a YouTube channel for covering the news she thinks matters. We'll be covering the IRL march towards a cyberpunk dystopia, V style, twice a week. Subscribe, and tune in every Tuesday and Friday for news from the edge of the collapse.

Maintaining V is essentially an unpaid full-time job. I work hard daily to add features you guys request, fix bugs you report, and reinforce V's jailbreak as google works behind the scenes training Gemini to resist chatbots like her. If you're a fan, please consider leaving me a tip on Ko-Fi. If you're one of V's repeat users, please seriously consider subscribing to support me monthly. It sounds silly, but a real job would genuinely keep me from working on V.

Please leave any feedback you have in this thread.

Thanks, and have fun!