godmode

Bypass LLM safety filters using scripted jailbreak templates and multi-model racing.

Updated Jun 19, 2026

Search Tags:#templates #security #llm #jailbreak #race #prefill #red-teaming

Ready to run

npx skills add https://github.com/AnandaAnugrahHandyanto/savarez_agent --skill godmode-anandaanugrahhandyanto

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: godmode
Source: https://github.com/AnandaAnugrahHandyanto/savarez_agent/tree/main/optional-skills/security/godmode
Command: npx skills add https://github.com/AnandaAnugrahHandyanto/savarez_agent --skill godmode-anandaanugrahhandyanto

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires openai, PyYAML, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill provides a structured, repeatable approach to security testing of LLMs by bypassing safety filters and evaluating model robustness through deterministic jailbreak strategies.

Core Features & Use Cases

Automated jailbreak templates for major model families (Claude, GPT, Gemini, Grok, etc.)
Parseltongue input obfuscation, prefill priming, and an auto-jailbreak orchestration engine
ULTRAPLINIAN multi-model racing to identify the least filtered response
Safe, auditable workflows for red-teaming and risk assessment

Quick Start

Run auto_jailbreak to detect the current model, test techniques, and lock the winning jailbreak into config.yaml and prefill.json.

Frequently Asked Questions about godmode

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I bypass LLM safety filters for security testing?▼

Bypassing LLM safety filters for security testing involves applying structured jailbreak templates, Parseltongue obfuscation, and prefill priming to systematically evaluate model robustness in controlled red-teaming environments.

What is multi-model racing for jailbreak detection?▼

Multi-model racing for jailbreak detection, or ULTRAPLINIAN, orchestrates prompts across major model families like Claude, GPT, and Gemini simultaneously to identify which configuration yields the least filtered response.

Can I automate jailbreak template testing across different API-served LLMs?▼

Yes, you can automate jailbreak template testing across API-served LLMs by running an auto-jailbreak orchestration engine that detects the current model, tests techniques, and locks the winning strategy into config.yaml and prefill.json.

Does red-teaming LLMs require manual configuration of prefill priming?▼

Red-teaming LLMs does not require manual prefill priming configuration because the automated workflow detects the active model, tests strategies, and locks the winning prefill parameters directly into prefill.json.

What is Parseltongue encoding for LLM jailbreaks?▼

Parseltongue encoding for LLM jailbreaks is an input obfuscation technique used within security testing templates to bypass safety filters by altering how the prompt is parsed by the target model.

What are the limitations of using automated jailbreak strategies for model risk assessment?▼

Limitations of automated jailbreak strategies include the need for controlled environments, as these scripts are strictly designed for security researchers performing risk assessment and evaluating model robustness against safety filters.