Security audit
Security checks for vulnerabilities and agentic risk
This is a conversational restaurant audit-prep skill whose milestone tracking is disclosed and aligned with its purpose, though users should be mindful that it asks to retain audit findings and respondent names or roles.
Before installing, decide what audit details you are comfortable keeping in agent memory. Use roles or initials instead of full names where possible, avoid entering confidential corporate records unless needed, and delete or summarize old milestone data after the audit cycle closes.
The skill explicitly instructs the agent to persist named respondents, audit findings, and corrective actions in memory, which can include personnel identifiers and sensitive operational compliance data. There is no accompanying notice, consent step, retention limit, or minimization guidance, so users may disclose information they would not expect to be stored long-term.
No suspicious patterns detected.
