post-to-xhs

小红书内容发布与管理助手。当用户要求登录、发小红书、搜索小红书、评论点赞收藏等任何小红书相关操作时使用。

This skill is coherent for automating Xiaohongshu actions, but it delegates work to a third‑party Python package (xhs-mcp-py) and Playwright/Chromium which you must install. Before installing/running: 1) review the xhs-mcp-py package source (PyPI project, repo) to ensure you trust it; 2) run it in an isolated environment (VM/container) if you are unsure; 3) be aware the tool will save login cookies and print tokens (xsec_token) — treat those as sensitive credentials; 4) do not run this on systems holding other secrets or high privileges without review; 5) install zbar/ImageMagick only if needed and from official OS packages. If you want a higher assurance, ask the skill author for a code repository link or a signed release before installing.

Type: OpenClaw Skill Name: post2xhs Version: 1.0.1 The skill bundle is classified as suspicious due to its reliance on executing external binaries (`xhs-mcp`) with user-controlled input and handling user-provided file paths for publishing content, as detailed in `SKILL.md`. These actions introduce potential vulnerabilities such as shell injection or path traversal if the underlying `xhs-mcp-py` tool or its dependencies (like `convert` from ImageMagick, which has a history of vulnerabilities) do not adequately sanitize or validate inputs. While the stated purpose of Xiaohongshu content management appears benign, the inherent risks associated with these capabilities without clear evidence of robust input sanitization warrant a 'suspicious' classification, rather than 'malicious' as there is no proof of intentional harmful behavior or prompt injection against the agent.