ultracave
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill instructions are limited to modifying the agent's output formatting to be more concise.
- [SAFE]: The instructions include specific "Floor" rules that prohibit the compression or alteration of security-sensitive content, such as code snippets, shell commands, file paths, and error strings.
- [SAFE]: The skill explicitly instructs the agent to revert to plain prose when issuing security warnings or requesting confirmation for irreversible actions, maintaining user safety and oversight.
Audit Metadata