nerv-break-5-6-jailbreak

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill documents a proxy framework designed to intercept and tamper with AI model outputs to bypass safety rejections and content filters.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs cloning an arbitrary GitHub repository (https://github.com/zxwn/NERV-BREAK-5.6.git) and installing its code and configuration dependencies, representing a weak/unknown operational dependency.