Security audit
Security checks for vulnerabilities and agentic risk
This skill creates a personal daily action brief from connected sources, with sensitive aggregation disclosed and no code, persistence, or destructive behavior found.
Install this only if you are comfortable with an assistant summarizing sensitive connected sources such as messages, email, calendar items, billing notices, weather context, and notes into one brief. Review the result in private contexts, and limit connected sources if you do not want financial or message content included.
The manifest description uses very broad invocation language such as daily planning, inbox triage, and proactive reminders across multiple sources. This can cause the skill to be selected for common user requests and aggregate sensitive personal data from email, messages, calendar, and notes more often than the user explicitly intended, increasing privacy and consent risk.
This skill is designed to aggregate reminders and alerts from highly sensitive channels including email, SMS/iMessage, billing notifications, calendar, weather, and task notes, but the description does not warn users that personal data across these sources may be combined. Cross-channel aggregation can expose sensitive relationships, schedules, finances, and communications in a single output, magnifying privacy harm if triggered unexpectedly or shown in the wrong context.
The trigger section designates several Chinese phrases as strong triggers, but the skill does not explain whether this language restriction is optional, user-configurable, or region-specific. This can create a language/locale policy concern because the skill appears to privilege a specific language without explicit opt-in or documented justification.
The example executive command is written in Chinese ("生成本周经营风险清单") while the rest of the file is in English, and the file does not state that outputs may vary by user language or that this mode is Chinese-specific. This can implicitly force a specific language in generated output without user opt-in, which matches the language/locale policy concern.
No suspicious patterns detected.
