Security audit
Security checks for vulnerabilities and agentic risk
This is a Chinese-language role-library integration with no bundled executable code, but it relies on external role definition files that should be trusted before use.
Install this only if you want Chinese-language expert-role behavior and you trust the separately installed /root/.openclaw/agency-agents-zh/ role files, because those external role definitions will shape how the agent responds. Review or control that local role directory if you need strict instruction provenance.
The skill is documented entirely in Chinese and presents Chinese-language interaction as the default without describing user choice, locale detection, or an opt-in mechanism. This can cause misunderstanding of instructions, unexpected language switching, and reduced transparency around what role behavior is being loaded, especially in multi-agent or safety-sensitive contexts.
The invocation examples are extremely broad and map closely to ordinary user requests, such as asking for a security engineer to review code or an architect to design a system. This can cause the skill to activate implicitly when the user did not clearly opt in, leading to unexpected persona loading, instruction interference, or overbroad delegation into other role files whose contents are not visible here.
No suspicious patterns detected.
