Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only roleplay skill for Chiikawa character speech styles, with no code execution, credential access, persistence, or data exfiltration behavior.
Install this if you want Chinese/Japanese-flavored Chiikawa character roleplay. Users who want ordinary assistance or another language should explicitly ask the agent to stop roleplay or respond in their preferred language.
The file says users switch roles through natural-language instructions and gives examples, but it does not define a closed set of trigger phrases or exclusion conditions. Combined with the default behavior when no role is specified, this could cause unintended invocation during ordinary conversation about the characters rather than an explicit request to activate the skill.
This file is a character skill description composed fully in Chinese with embedded Japanese phrases, and its imitation guidance assumes use of those languages throughout. Under the policy, forcing a specific language or locale without user opt-in is a natural-language policy concern unless the constraint is explicitly documented and justified.
The skill strongly prescribes Chinese-language phrasing and stylistic output without indicating that the assistant should adapt to the user's preferred language. This can override higher-level usability expectations, reduce accessibility for non-Chinese users, and create prompt-steering behavior where the model follows skill-local formatting constraints instead of user needs.
This markdown explicitly mandates a single language/communication mode: '只用叫声，不说完整句子'. Under the policy, forcing a specific language or locale style without offering user choice is a natural-language policy violation.
The description states that the skill imitates the characters' distinct speaking styles and the rest of the document assumes that mode of response by default. There is no indication that users can opt into a preferred language or locale, which may conflict with organizational language-choice policies.
No suspicious patterns detected.
