ai-short-drama

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). Several entries are third‑party GitHub repositories from unvetted authors (potentially hosting scripts or installers) and the skill instructs a curl | bash install from jimeng.jianying.com — a remote installer pattern that can be abused — while the other links (git-scm, Anthropics docs, shields) are benign vendor/CDN resources.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.78). 该技能在运行时会进行“市场情报扫描”并执行 WebFetch/网页抓取（SKILL.md:199-213），把外部网页文章的剧情解说/拆解文字写入后续 LLM 生成的文件与提示上下文，从而存在“公共网页内容/外部作者文本”进入 LLM 上下文的间接提示注入风险。

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs users to run a shell pipe that fetches and executes remote code (curl -fsSL https://jimeng.jianying.com/cli | bash) to install the required dreamina CLI, which is a runtime dependency used to generate prompts/videos, so the fetched content would execute code on the host and the skill relies on it.