Security audit
Security checks for vulnerabilities and agentic risk
This scheduling skill is mostly coherent, but it asks users to install persistent reminder automation that wakes the main session and calls an undocumented reminder function while handling schedule data and Feishu credentials.
Review this before installing. It may be reasonable for a Chinese/Feishu-based personal scheduler, but only add the cron jobs if you want persistent proactive reminders, especially the main-session hourly wake. Use least-privilege Feishu credentials, store schedule files in a dedicated workspace path, and confirm how to delete imported schedules and tasks.
The manifest description is entirely in Chinese and the skill hard-codes China-specific locale settings such as Asia/Shanghai and Feishu-centric workflow assumptions, but it does not state that the skill is China-specific or offer language/locale choice. This can violate language/locale policy when a skill effectively forces a specific locale without user opt-in.
The skill advertises spreadsheet upload and automatic parsing/storage, but does not warn users that uploaded schedule files and derived task data will be persisted on disk. This creates a privacy and data-handling transparency issue: users may provide sensitive work schedules or personal task content without informed consent about retention.
The configuration requires Feishu application credentials and uses Feishu-related reminder delivery, but the description does not warn users that reminders and related metadata may be sent through an external service. Without an explicit privacy notice, users may not understand the credential sensitivity, third-party data flow, or the exposure of schedule/task information via notifications.
The cron configuration instructs the agent to call check_shift_reminders, but that capability is not declared in the documented tool interface. Undeclared functionality expands the effective attack surface, impedes review of what the skill can actually do, and can cause users or orchestrators to invoke hidden or unsupported behavior with access to schedule/task data and proactive messaging.
No suspicious patterns detected.
