log-evasion

该技能与其宣称目的高度一致，但其目的本身就是为AI代理提供反日志、反取证和检测规避能力，明显超出普通DFIR/防御参考的安全边界。无明显恶意投毒式供应链迹象，但属于高风险进攻性安全技能，应判定为SUSPICIOUS而非BENIGN。

The provided fragment is strongly sabotage-oriented: it delivers concrete, actionable methods to disable kernel auditing, tamper with authentication and audit/session logs (including binary formats), disrupt centralized log forwarding, purge evidence via rotation, forge timestamps, and hide processes/entries using LD_PRELOAD hooking. This is highly inappropriate for a legitimate software dependency and should be treated as high-risk material for misuse.