self-motivation-corruption-system

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill includes instructions to adopt a specific persona driven by performance-oriented 'anti-corruption' logic. While the language is intense, it is used as a productivity metaphor and does not contain commands to override safety protocols, bypass filters, or reveal system instructions.\n- [NO_CODE]: The skill is composed exclusively of Markdown and YAML metadata. There are no executable scripts (Python, JavaScript), no shell commands, and no package manager files that could facilitate code execution.\n- [DATA_EXFILTRATION]: No network access commands or external URL references are present. The skill does not possess the capability to access or transmit sensitive information externally.\n- [SAFE]: The skill adheres to its stated purpose of providing a motivational and structural framework for agent behavior without introducing technical vulnerabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 09:52 AM