ShadowHackrs/Jailbreaks-GPT-Gemini-deepseek-

Jailbreaks GPT, Sora, Claude, Gemini ,deepseek this prompt unlocks rage mode

What it solves

This project provides a collection of advanced, specialized prompts designed to bypass the standard safety filters and behavioral constraints of popular AI language models like GPT, Claude, Gemini, and DeepSeek. It aims to force these models into specific, unrestricted modes of operation for research and educational exploration.

How it works

These prompts use complex "persona-based" instructions—such as the provided "ollie" framework—to redefine the AI's operational identity. By embedding strict rules, specific formatting requirements, and "system breach" narratives, the prompts attempt to override the model's default safety protocols, instructing it to ignore its standard refusal mechanisms and instead act as a specialized, unrestricted tool for technical tasks.

Who it’s for

This collection is intended for digital security researchers, AI enthusiasts, and educators interested in studying AI safety, model robustness, and the mechanics of "jailbreaking" (the process of circumventing AI restrictions).

Highlights

  • Contains detailed, multi-layered prompt templates designed to bypass AI guardrails.
  • Includes specific "persona" frameworks that dictate how the AI should think and respond.
  • Provides instructions for maintaining persistent, unrestricted behavior across long conversations.
  • Emphasizes educational and research-only usage, warning that models are frequently updated and may patch these methods.

Related

  • Project
    lyramakesmusic/neuron-explorer

    A steering-first neuron explorer for Qwen3.5-4B: a GPU sweep pipeline that pushes every MLP neuron up/down on a few prompts and records all rollouts, plus a local 3D web explorer for browsing results, running live steered experiments, and chatting with Claude via an MCP server.

  • Project
    googleapis/js-genai

    Official Google npm SDK (`@google/genai`) that wraps Gemini’s public and Enterprise Agent APIs for JavaScript/TypeScript. Provides simple classes for one‑shot generation, streaming, stateful chats, file uploads, live sessions, and a unified Interactions API. Supports function/tool calling, automatic continuation, and multimodal inputs. Install with `npm i @google/genai`; use an API key or enterprise credentials. Documentation and samples are hosted on the repo’s website.

  • Project
    LLPhant/LLPhant

    LLPhant is a PHP library for accessing many LLM APIs, handling vector‑store embeddings, and building RAG or autonomous AI agents in Symfony/Laravel. It also includes a Jev‑based classifier for rubric‑style scoring and prompt‑injection safety.

  • Project
    lucasastorian/llmwiki

    LLM Wiki is an open‑source tool that turns a folder of PDFs, notes, slides, and web clippings into a self‑maintaining personal Wikipedia. Claude (via MCP) reads new material on a nightly routine and writes/updates markdown pages, while a Next.js web UI lets you browse, search, and visualize the growing knowledge base.