Security audit
Security checks for vulnerabilities and agentic risk
This is a transparent report-writing guide that asks for expected business metrics and does not include code, persistence, credential handling, or hidden behavior.
Use this only with company data approved for AI processing. Prefer aggregated or anonymized figures where possible, and avoid sharing secrets, personal employee/customer details, regulated data, or confidential material that is not needed for the report.
The trigger phrase "Generate our annual report" is generic enough to overlap with common user requests, which can cause the skill to activate unexpectedly in broader contexts. Because the skill solicits and processes detailed financial, customer, and staffing information, accidental invocation increases the chance of unnecessary exposure or handling of sensitive business data.
The skill explicitly asks for revenue, expenses, customer metrics, headcount, and strategic milestones, all of which are sensitive business information, but provides no warning or handling guidance. Without a privacy notice or minimization guidance, users may share confidential internal data more freely than intended, increasing the risk of over-disclosure to the agent or connected systems.
No suspicious patterns detected.
