XHS Viral Content Factory

小红书爆款图文全自动生产工厂。支持从 PDF、Markdown 或文件夹提取内容，自动匹配治愈、干货、反直觉、视觉流 4 大模式。

What to consider before installing: - Data exfiltration risk: The script sends the contents of whatever files/folders you point it to a remote LLM. Do NOT pass folders that contain secrets, config files, SSH keys, environment files, or other sensitive data. - Trusted provider: Ensure your LLM_API_KEY is for a provider you trust and understand the provider's data retention/privacy policy. If you must protect data, run on a provider that guarantees no retention or use a local/self-hosted model. - Inspect the files: Review generate_xhs.py and SKILL.md yourself (or have someone you trust do it). The pre-scan flagged hidden unicode control characters in SKILL.md — open the file in a hex-capable editor or use a control-character filter to confirm nothing malicious is hidden. - Run in isolation: If you try it, run in a disposable environment/container and with an API key that has limited scope or billing limits, not a high-privilege production key. - Usage hygiene: Do not export your main system-wide API key on shared machines. Avoid pointing the tool at system-level directories (e.g., /etc, ~, or project roots) — only point it to curated content you intend to send. If you are uncomfortable with these risks, do not install. If you proceed, consider modifying the script to explicitly whitelist file types/paths and to prompt for confirmation before sending any content.

Type: OpenClaw Skill Name: xhs-viral-factory Version: 1.1.3 The skill is a legitimate tool designed to automate the creation of Xiaohongshu (XHS) social media posts by processing local text/markdown files through a user-configured LLM API (e.g., OpenAI or DeepSeek). The core logic in `scripts/generate_xhs.py` correctly handles environment variables for secrets, implements reasonable file-reading constraints, and includes a transparent security disclosure regarding data transmission to the chosen LLM provider. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found.