revolver-review
Pass
Audited by Gen Agent Trust Hub on Jul 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates as a high-level process guide ('agent-guide') and does not include any code, external dependencies, or specific tool invocations that could pose a security risk.
- [PROMPT_INJECTION]: The instructions for subagents to adopt an 'adversarial discipline' (e.g., 'Refute by default') are context-specific directives for code review quality and do not attempt to bypass the AI agent's core safety filters or ethical guidelines.
- [DATA_EXFILTRATION]: No patterns for accessing sensitive files (such as .env, .ssh, or cloud credentials) or exfiltrating data to external domains were detected.
- [REMOTE_CODE_EXECUTION]: The skill does not contain any mechanisms for downloading or executing remote code, nor does it use dynamic execution patterns like eval or subprocess for untrusted input.
- [SAFE]: The 'revolver mechanism' described is a logical workflow for managing agent interactions and does not introduce persistence mechanisms or privilege escalation risks.
Audit Metadata