writing-assistant

SUSPICIOUS. The core writing workflow is coherent, but the skill's footprint includes autonomous public posting, browser-cookie-backed platform access, external content ingestion with write/exec powers, and transitive trust in bundled companion skills. This is not clearly malicious, yet it is broader and riskier than a normal writing assistant and should be treated as a high-risk agent skill.

该技能的文案生成部分基本符合用途，但其实际 footprint 明显扩大到安装第三方二进制、控制本地 GUI 登录流程，并代表用户执行小红书公开账号操作。数据流大多与声明用途一致，未见明确的凭据窃取或外传端点，因此更像高风险自动化技能而非确认恶意。总体应判为 SUSPICIOUS：能力范围大、真实世界动作强、安装信任一般。