sudongpo-perspective

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides instructions for the agent to adopt a specific persona using well-defined emotional and philosophical frameworks. No prompt injection, safety bypass attempts, or instructions to ignore system rules were found.- [EXTERNAL_DOWNLOADS]: The skill installation instructions reference npx skills add nuwa-skills/sudongpo-skill. This package belongs to the vendor's namespace and is consistent with the provided author context. All external links in the documentation point to research sources or the developer's repositories.- [DATA_EXFILTRATION]: No evidence of hardcoded credentials, sensitive file access (such as .ssh or .env), or unauthorized network operations was found in any of the skill files.- [COMMAND_EXECUTION]: The skill does not contain any shell commands, scripts, or privilege escalation patterns. The persona is strictly conversational and does not invoke external system tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 02:19 PM
Security Audit — agent-trust-hub — sudongpo-perspective