Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed image-prompt generator that sends user-entered descriptions to the local OpenClaw Gateway and does not show hidden persistence or destructive behavior.
Before installing, confirm that you trust the model provider configured behind your OpenClaw Gateway and avoid entering confidential creative briefs or client material unless that provider is acceptable. The skill would benefit from narrower trigger phrases, but the reviewed artifacts are purpose-aligned and disclosed.
The trigger phrase '生成提示词' is overly generic and can match many non-image-related requests, causing the skill to activate outside its intended scope. This increases the risk of prompt hijacking or unintended routing, where unrelated user requests are handled by this skill and their content is forwarded to the configured model provider.
The pattern '用[风格]风格生成' is ambiguous because it lacks any image-specific constraint and could match broad content-generation requests such as writing, code, or document creation. In this skill, that ambiguity is more dangerous because matched requests may be transformed and sent to the model backend, leading to unintended invocation and possible disclosure of user input to the downstream provider.
The natural-language instruction explicitly requires output to be in Chinese, and the UI also hardcodes Chinese-language behavior, without offering the user a language selection or documenting a justified region-specific constraint. This is a locale policy issue because it imposes a specific language regardless of user preference.
The app sends raw user-entered content to a remote LLM API via fetch() without any visible disclosure, consent flow, or privacy notice in the UI. In a prompt-generation tool, users may paste proprietary creative concepts, client materials, or sensitive text, so silent transmission creates a genuine privacy and data-handling risk even if there is no direct code execution issue.
The description is written entirely in Chinese, which imposes a specific language/locale in the skill metadata without indicating that users can choose another language. This can violate language/locale policy where skills should not force a language unless clearly justified or optional.
No suspicious patterns detected.
