qianxuesen-perspective

Pass

Audited by Gen Agent Trust Hub on Apr 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is composed of documentation, role-play guidelines, and theoretical frameworks. It does not contain any functional code or scripts.
  • [PROMPT_INJECTION]: The instructions focus on persona adoption and systematic thinking methodologies. There are no attempts to override safety filters, extract system prompts, or use deceptive role-play to bypass restrictions. The skill includes an 'Honest Boundary' section that explicitly guides the agent to handle historical controversies with factual accuracy.
  • [COMMAND_EXECUTION]: No shell commands, system calls, or administrative operations are requested or defined in the skill files.
  • [DATA_EXFILTRATION]: The skill does not define any network tools and does not attempt to access sensitive file paths such as SSH keys or environment secrets.
  • [EXTERNAL_DOWNLOADS]: No external scripts or remote resources are downloaded or executed. The installation command mentioned in the README is for the skill platform itself and is not a dependency of the skill's logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 13, 2026, 09:46 AM
Security Audit — agent-trust-hub — qianxuesen-perspective