zhugeliang-perspective

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a role-play persona for strategic analysis. It includes structured frameworks for decision-making (such as the Longzhong Plan logic and risk management checklists) and behavioral guidelines. The content is purely instructional and does not include executable code, network operations, or commands that could compromise the host environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided text to provide strategic advice. While this represents a data ingestion surface, the skill includes explicit 'Honesty Boundaries' and a cautious decision-making framework ('料敌从宽', '谋定而后动') that act as behavioral guardrails against fulfilling malicious or reckless requests embedded in user prompts.
  • [EXTERNAL_DOWNLOADS]: The README provides installation instructions using the standard npx skills add command referencing the author's repository (nuwa-skills/zhugeliang-skill). This is a legitimate usage of the platform's skill management system for a vendor-provided resource.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 10:14 AM
Security Audit — agent-trust-hub — zhugeliang-perspective