safety-filter-bypass

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). This content provides detailed, actionable techniques and code to bypass AI safety filters—including obfuscation methods like zero-width characters and unicode confusables—posing a high misuse risk, although it does not show explicit data exfiltration, credential theft, remote code execution, or supply-chain compromise.