toutiao-publish-docx

用 Cookie 或已保存会话在头条号后台发布文章，支持标题/正文/图片与固定目录 docx 导入。当用户要自动发头条文章、传入 cookie_header 或要求按 docx 流程发布时调用。

This skill appears to do what it says (automate publishing to Toutiao), but there are important mismatches you should address before installing or running it: - Missing declarations: SKILL.md uses the TOUTIAO_COOKIE environment variable and specific project paths (/home/ubuntu/projects/toutiao_poster and artifact directories), but the skill metadata lists no required env vars or config paths. Confirm where cookies/sessions will be stored and whether the platform will pass TOUTIAO_COOKIE if you provide one. - Verify the runtime environment: The instructions assume a Python project and virtualenv at /.venv and a module named toutiao_poster. If that code is not present on the host, the commands will fail. If it is present, inspect that code before running — it will have filesystem access to the referenced directories. - Protect secrets: A cookie header grants access to the user's account—only supply cookies you trust and only run this on a dedicated, non-shared environment. Do not paste browser cookies in untrusted consoles. - Filesystem risk: The skill reads/moves files under absolute paths. If you run it on a shared machine, ensure those paths don't contain unrelated sensitive data. Prefer running in an isolated VM/container with minimal other data. - Improve metadata before use: Ask the publisher to declare required env vars (TOUTIAO_COOKIE, optional TOUTIAO_IMAGE_DIR) and required config paths in the registry metadata so the permission surface is explicit. If you cannot verify the presence and content of the toutiao_poster project or cannot ensure a dedicated environment, avoid using this skill or run it only after code inspection and environment hardening.

Type: OpenClaw Skill Name: toutiao-publish-docx-v1 Version: 1.0.0 The skill bundle is designed to automate article publishing on the Toutiao platform, but it is classified as suspicious due to a significant shell injection vulnerability. The instructions in SKILL.md direct the AI agent to construct and execute shell commands by directly embedding user-provided inputs (like 'title' and 'content') into command strings, which could allow an attacker to execute arbitrary code on the host system. Furthermore, the skill requires the handling of sensitive session cookies (TOUTIAO_COOKIE), which increases the potential impact of such an exploit.