知识星球自动化技能。支持发帖、回帖、浏览帖子、检查通知、自动回帖、自动发帖。自动读取 ~/.xfg-zsxq/groups.json 配置。支持定时任务自动检查通知和被圈提醒。回复功能使用浏览器自动化(Playwright MCP)。触发词:'知识星球'、'发帖'、'回帖'、'zsxq'、'检查通知'、'浏览帖子...
Security Analysis
high confidenceThe skill's code, instructions, and requested access are consistent with its stated purpose (automating 知识星球 posts/notifications); it primarily needs the user's knowledge-site cookie and local config files and does not request unrelated credentials or system-wide privileges.
The name/description describe automation for 知识星球 (post, reply, check notifications, cron). The repository includes scripts that implement those features and read/write ~/.xfg-zsxq/groups.json. Required capabilities (HTTP requests to api.zsxq.com, cookie-based auth, optional image upload to qiniu) match the stated purpose. There is a minor doc mismatch: SKILL.md mentions Playwright MCP for browser automation while package.json lists puppeteer as a dependency; this is a documentation/implementation inconsistency but does not indicate malicious behavior.
SKILL.md and scripts instruct the agent/user to obtain and store the full site Cookie and to run local Node scripts. The runtime instructions explicitly reference only the knowledge-planet APIs, the local config path (~/.xfg-zsxq/groups.json), and the OpenClaw browser tooling for interactive automation. There are no instructions to read unrelated system files or to exfiltrate data to unexpected endpoints beyond api.zsxq.com and qiniu (used for image upload).
This is instruction- and script-based with no platform install spec. A package.json declares 'puppeteer' (heavy dependency) but there is no automated install step described; users will need to npm install if they want browser automation. Because no install script runs automatically, there's lower installer risk, but users should be aware that dependencies (puppeteer) are required for automated browser actions.
The skill requires the user's 知识星球 cookie (zsxq_access_token) stored in ~/.xfg-zsxq/groups.json and uses it to call api.zsxq.com. Requesting the cookie is proportionate to performing authenticated post/reply operations. The code stores configuration with file mode 0o600 (README claims permission 600), which is appropriate for sensitive tokens. No unrelated environment variables or unrelated service credentials are requested.
always:false and no special OS-wide privileges are requested. Scripts read/write files under the user's home (~/.xfg-zsxq and cron-config.json) which is expected for a per-user automation tool. The cron-setup script writes a cron-config JSON and prints OpenClaw CLI commands but does not autonomously register system cron jobs or modify other skills' configurations.
Guidance
This skill is coherent with its stated purpose but it handles sensitive data: it asks you to paste your full 知识星球 Cookie (contains zsxq_access_token). Before installing or running: 1) Review the scripts yourself and only store cookies you trust this tool with; the tool saves them under ~/.xfg-zsxq/groups.json (recommend file mode 600 and only use on accounts you control). 2) Be aware browser automation requires installing puppeteer (package.json) or using the platform's browser actions; follow README instructions and install dependencies in an isolated environment if possible. 3) Image uploads go to qiniu (upload-z1.qiniup.com) which is expected for the service but note third-party storage. 4) If you expect the agent to run autonomously, understand scheduled task instructions only produce config and CLI commands — the script doesn't silently create system cron jobs. 5) Minor docs mismatch: SKILL.md mentions Playwright while package.json lists puppeteer; confirm which automation tool you'll run. If any of this is unexpected, do not provide your cookie or run the scripts until you are comfortable.
Latest Release
v2.3.0
- 新增详细的 SKILL.md 文档,完善了各项功能使用说明。 - 支持发帖、回帖、浏览帖子、检查通知、自动回帖和定时任务等自动化流程。 - 回帖功能现采用浏览器自动化(Playwright MCP)提升兼容性。 - 丰富常见问题、API 状态、项目结构和 Cookie 配置说明,便于使用和排查。 - 触发词与路由功能增强,更好适配自动响应和用户操作习惯。
More by @fuzhengwei
Published by @fuzhengwei on ClawHub