byted-kickart-viral-replicator

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). High-risk: the docs explicitly instruct agents to ask users to paste ACCESS_KEY_ID / SECRET_ACCESS_KEY into chat and export them to environment variables (credential-exfiltration/social-engineering), and the code logs request headers/bodies (including Authorization) to local logs, increasing leakage risk.

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

• Third-party content exposure detected (medium risk: 0.30). 在“素材分析/爆款裂变”流程中，用户可提供抖店/抖音商品链接，随后运行 scripts/material.py 通过 requests.head + 后端拉取并解析该页面/商品所含文本（如 material.py 收集到的 url 等结果）进入后续 LLM/分析参数，属于外部用户可提交内容且先于“选择特定项目”即被摄取的间接提示注入风险（链接文本被后端返回后也可能进入结果文件再被后续步骤使用）。

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

• Potentially malicious external URL detected (high risk: 0.80). The guide in references/套餐开通指南.md instructs running "pip install -r ./scripts/requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple" which will fetch and install remote Python packages at runtime (executing code from that external mirror), and the skill treats these dependencies as required preconditions.