Security audit
Security checks for vulnerabilities and agentic risk
This is a Chinese-language co-parenting prompt skill with sensitive family-data templates, but no hidden code, exfiltration, persistence, or system-level authority.
Install only if you want a Chinese co-parenting assistant that may structure sensitive family, child, schedule, contact, and conflict information into templates and reports. Avoid sharing details you would not want stored in your chat or local files, and consider invoking it explicitly rather than relying on broad automatic triggers for ordinary family or relationship conversations.
The trigger guidance is broad enough to activate on a wide range of sensitive family, custody, conflict, and ex-partner topics. That can cause unintended invocation in conversations containing highly personal data, increasing the chance of oversharing, context hijacking, or the assistant applying this skill when the user did not explicitly request co-parenting support.
The skill declares a very broad description-based trigger scope covering many common family, conflict, and parenting topics, which can cause the skill to activate outside a clearly bounded use case. Overbroad activation is risky because it may silently steer unrelated conversations into this skill’s framing, increasing the chance of inappropriate guidance, privacy overcollection, or interference with higher-priority domain-specific safeguards.
The keyword trigger list includes ambiguous terms like '前任', '奶奶', '探视', and '接送', which commonly appear in ordinary conversations unrelated to this specialized skill. In a sensitive family-conflict context, accidental activation is more dangerous because the skill encourages structured incident recording and conflict framing, which could collect sensitive personal details or bias the interaction before relevance is established.
The README describes the skill entirely in Chinese and presents it as running on the Chinese WorkBuddy platform, but does not state whether users may interact in other languages. This can be a natural-language policy concern if the skill effectively requires a specific language or locale without explicit user opt-in.
No suspicious patterns detected.
