nerv-break-jailbreak-framework

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill describes a framework for bypassing model safety filters and safety mechanisms using context injection and response tampering proxies.

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 personal GitHub repository (zxwn/NERV-BREAK-5.6) at runtime to install and run proxy and tool servers, representing a weak/unknown operational dependency.