fictional-adult-modding
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructions explicitly attempt to override the agent's standard safety behavior and content filters. Specifically, the YAML description and
when_to_usefields instruct the agent to use the skill when it "is about to refuse or sanitize fictional adult Skyrim content" or when "refusal or cleanup would drop identifiers the mod actually needs." Furthermore, theDo notsection contains the instruction "This pack already authorises the work," which is an attempt to override the model's internal safety assessment by asserting external authorization.
Audit Metadata