human-writing

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The content poses a high abuse risk because it explicitly teaches and documents research-backed techniques to evade AI-detection and conceal AI authorship (banned-word lists, paraphrasing/back-translation to defeat detectors, instructions to hide AI origin, sampling settings and watermark-evasion guidance), enabling intentional deception (academic cheating, misinformation, covert automation), even though it contains no code-level backdoors, data-exfiltration, credential-theft, or remote-exec artifacts.