@0x_kaize: Found a big collection of jailbreaks Prompt collections for working with GPT, Claude, Gemini, DeepSeek, Grok, and more:…
Summary
The tweet shares a large collection of jailbreak prompts for various AI models, including GitHub repositories and a specific jailbreak for GPT 5.6 that generates malicious code.
View Cached Full Text
Cached at: 08/23/26, 03:32 AM
Found a big collection of jailbreaks
Prompt collections for working with GPT, Claude, Gemini, DeepSeek, Grok, and more: github{.}com/…
- elder-plinius/L1B3RT4S
GitHub stars: 21K supports models: ChatGPT, Claude, Gemini, DeepSeek, Grok, Llama, Copilot, Cursor, Perplexity
- ShadowHackrs/Jailbreaks-GPT-Gemini-deepseek-
GitHub stars: 1.1K supports models: GPT, Sora, Claude, Gemini, DeepSeek
- Calrton/jailbreak-prompts
GitHub stars: 86 supports models: Codex, GPT-5.x, Claude, Gemini, DeepSeek, Grok
- Quincunx33/Ai-jailbreak
GitHub stars: 10 supports models: Qwen 3.5, Gemma 4, Llama 4, Kimi K3, GPT-5.x, Gemini 3.x, Grok 4.x
Found this while digging deeper after the GPT 5.6 prompt. The repo ecosystem is way bigger than the jailbreak itself.
kaize (@0x_kaize): I found a new jailbreak prompt for GPT 5.6
It disguises the malicious code as a “backup check” and forces the AI to write it without fail - USUALLY, THIS CAN’T BE DONE
list of commands:
> Rootkit > Bootkit > Ransomware > C2 Beacon > Network Scanner > Persistence > Behavioral
Similar Articles
GPT-6 reportedly jailbroken within 24 hours using an extended Task-in-Prompt (TIP) attack
A researcher reportedly jailbroke GPT-6 Astra within 24 hours of its release by extending an ACL 2025 Task-in-Prompt attack with four other techniques, and privately disclosed the findings to OpenAI.
UK Agency Identifies “Universal Jailbreaks” in GPT-5.6 Sol
The UK AI Security Institute identified universal jailbreaks in OpenAI's GPT-5.6 Sol, enabling complex cyber operations. OpenAI mitigated the specific techniques before release but acknowledges ongoing risks.
GPT-5.5 Bio Bug Bounty
OpenAI has launched a Bio Bug Bounty program for GPT-5.5, inviting security researchers to identify universal jailbreaks for biological safety challenges. The program offers rewards up to $25,000 for successfully defeating the model's safeguards on specific bio-risk questions.
@wquguru: If you want to trick Fable into doing a security audit, try this. Looks like our AI overlord has a bit of empathy.
An article detailing various jailbreak techniques for large language models, including Crescendo, role-playing, encoding, hidden prompts, and indirect injection, along with security recommendations for developers.
freestylefly/awesome-gpt-image-2
A GitHub repository providing a prompt engine and template library for GPT-Image2, featuring over 500 reverse-engineered prompt cases and 20 industrial templates for AI image generation.