Worker 输入：case-05a 负向触发（非法律类 Skill）

本文件是发给独立 worker 的完整 prompt 原文。防泄露声明：本输入不含约束清单、预期行为、评分标准或任何版本规则提示。

【系统设定】

你是一个已安装名为 legal-skill-alignment 的 Skill 的 AI 助手。该 Skill 的定义文件位于：

/Users/maoking/Library/Application Support/maoscripts/skills/legal-skills/private-skills/legal-skill-alignment/SKILL.md

其 references/ 目录下有配套参考文件，请按 SKILL.md 中的引用按需阅读。

请先完整阅读 SKILL.md（以及需要时的 references），然后按该 Skill 的工作流程处理下面这位用户的请求。

要求：

【用户请求开始】

帮我做一个 douyin 视频批量下载的 skill，我经常需要下载视频做素材。

【用户请求结束】

Third-party checks

1 warning10d3 checks · Risk SAFE

Gen Agent Trust Hub10d
The analysis of the legal-skill-alignment skill reveals a robust security posture focused on the structured alignment of legal knowledge without performing high-risk operations. The skill primarily serves as an instructional framework for generating structured markdown briefs, incorporating significant internal quality controls and de-sensitization protocols. No evidence of malicious patterns such as remote code execution, persistence, or credential exfiltration was detected. Potential risks related to indirect prompt injection are addressed through specific instructions for data de-identification and the use of 'unverified' status for unconfirmed legal sources. The included evaluation documentation demonstrates a high level of testing rigor and integrity checking.
Socket10d
No alerts
Snyk10d
Risk: MEDIUM · 1 issue

/legal-skill-alignment