Temp Notification Reader

读取 macOS 通知中心，导出到 memory/computer_io/notification/ 目录。让 OpenClaw 了解你的工作动态。

This skill appears to do what it says: read your local macOS notification database and write local summaries for OpenClaw. Key things to consider before installing: - Privacy: the tool reads notification titles and bodies (potentially sensitive). Review the included Python and shell scripts yourself to confirm you are comfortable with what will be collected and where it is stored (~/.openclaw/workspace/memory by default). - macOS permissions: to access the notification DB you'll likely need to grant notification access and Full Disk Access to python3 (a broad privilege). Only grant these if you trust the code and its source. - Source verification: SKILL.md and README reference a GitHub repo (https://github.com/gift-is-coding/...), but the registry metadata shows unknown source/homepage. Confirm the repository origin and review commit history before cloning or running scripts. - Cron / scheduling: adding a recurring job will repeatedly collect notifications. Test the export scripts manually first to inspect outputs, then add scheduling if desired. - Running safely: run the scripts manually in a controlled environment (not as root), consider using a dedicated Python virtualenv, and inspect the files created in /tmp and the OpenClaw memory directory. Avoid a global npm install unless you verify the published package. If you want lower-risk testing, run read_notifications_visual.py or read_notifications.py with a short lookback and output to a local temp file to inspect results before enabling persistent scheduling or granting broad OS permissions.

Type: OpenClaw Skill Name: temp-notification-reader Version: 1.0.0 The skill is designed to read the macOS notification database (db2/db) via read_notifications.py, which requires 'Full Disk Access' and provides access to sensitive private communications, including messages from Teams, Outlook, and WeChat. While the logic appears to stay local and aligns with the stated purpose of generating work summaries, the capability to programmatically capture all system notifications is a high-risk behavior that could be used to intercept 2FA codes, private messages, or other sensitive data appearing in notifications.