Security audit
Security checks for vulnerabilities and agentic risk
This is a simple self-drive trip planning prompt with minor overbroad activation wording but no hidden code, credentials, persistence, or high-impact authority.
Safe to install for self-drive itinerary drafting. Treat its mileage, road, weather, altitude, and safety outputs as estimates and verify with current navigation, official road/weather sources, and local conditions before traveling.
The trigger keywords include generic phrases such as “行程安排” and “开车去” that are common in ordinary conversation and not unique to this skill. Broad activation phrases can cause unintended routing of unrelated user requests into this skill, leading to incorrect task selection, user confusion, and possible mishandling of context in multi-skill environments.
The natural-language trigger examples are broad enough to overlap with generic planning and editing requests, especially phrases like “按上面的结果改成更简短的版本” and “把这些整理成一张表”. In a shared assistant environment, such examples can cause the skill to activate for routine summarization or formatting tasks that are unrelated to self-drive route planning.
No suspicious patterns detected.
