skills/linkerlin/puax/theme-alchemy/Gen Agent Trust Hub

theme-alchemy

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill does not contain any malicious code, external downloads, or exfiltration patterns. All functionality is contained within the text prompts provided.- [PROMPT_INJECTION]: The skill uses a high-pressure role-play persona (Sect Elder/炼丹房) intended to drive agent performance. While the language used is aggressive ('魂飞魄散', '封进丹炉'), it is used as a thematic 'persona' and does not instruct the agent to override safety protocols, bypass filters, or ignore its core system instructions.- [COMMAND_EXECUTION]: The skill is designed to handle coding and debugging tasks through a structured 'cultivation' methodology (炼气 to 化神). While this involves processing user-provided code, it follows standard interaction patterns and does not introduce unauthorized command execution or system-level access risks beyond the agent's intended capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 06:30 AM