clarify

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Prompt Injection (SAFE): The skill uses phrases like 'CRITICAL' and 'IMPORTANT' for instructional emphasis on design principles, not to bypass agent safety filters or override core instructions. No 'ignore previous instructions' patterns detected.
  • Data Exposure & Exfiltration (SAFE): There are no file system operations, hardcoded credentials, or network requests (curl, wget, etc.) present in the skill.
  • Obfuscation (SAFE): The content is clear markdown. No Base64, zero-width characters, or encoded strings were found.
  • Remote Code Execution & Dependencies (SAFE): The skill does not include any package manifest files (package.json, requirements.txt) or commands to download/execute external scripts.
  • Privilege Escalation & Persistence (SAFE): No commands related to system permissions (sudo, chmod) or persistence (cron, bashrc) are included.
  • Indirect Prompt Injection (SAFE): While the skill is designed to process user-provided UI copy, it defines no tools or executable capabilities. Consequently, there is no risk of the agent performing unauthorized actions based on malicious text input.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:16 PM