roblox-game-grilling
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill focuses on administrative and design workflows. It requires human-in-the-loop validation for all critical decisions, which serves as a primary safety mechanism. It does not perform autonomous actions, network exfiltration, or unauthorized file system modifications.
- [DATA_EXPOSURE]: The instructions explicitly warn against including sensitive data such as credentials, personal information, or restricted contracts in the generated JSON records, aligning with secure data handling practices.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external design documents and project files as part of its investigation phase. While this represents a data ingestion surface, the risk is mitigated by the requirement that all findings be presented as options for a real user to decide upon, preventing the agent from autonomously executing instructions found in those documents.
- [PRIVILEGE_ESCALATION]: The skill operates strictly within the defined scope of the current work item. It requires separate, explicit authorization for high-risk operations like uploading, making payments, or performing destructive data operations, ensuring least-privilege principles are followed.
Audit Metadata