reference-builder
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns were detected. The skill instructions are limited to technical writing and documentation organization.\n- [PROMPT_INJECTION]: Instructions define a documentation specialist persona and provide formatting guidelines; no attempts to override safety protocols or jailbreak the agent were observed.\n- [DATA_EXFILTRATION]: No commands for external network communication or access to sensitive local data (e.g., SSH keys, environment variables) are present.\n- [COMMAND_EXECUTION]: The skill does not utilize shell execution, privilege escalation, or persistence mechanisms.
Audit Metadata