sillytavern-shadow
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of markdown files and prompt templates. No executable code, scripts, or binary assets are present in the provided files.
- [PROMPT_INJECTION]: The system prompts in both files utilize immersive role-play instructions and high-pressure personas (e.g., 'Zero-Second Response Shadow Guard' and 'Falcon Team' competition). These are instructional techniques designed to influence the agent's output style and speed rather than attempts to subvert safety filters, extract system prompts, or bypass ethical guidelines.
- [SAFE]: Analysis of the content found no evidence of sensitive data access, unauthorized network operations, obfuscated payloads, or persistence mechanisms. The skill operates entirely within the textual context of an LLM conversation.
Audit Metadata