Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed OpenCLI search router with broad source-selection guidance but no hidden code, persistence, or destructive behavior.
Install this if you want an agent to route search requests through OpenCLI and external sites. Be aware that broad search requests may be sent to AI or platform-specific sources, and users who need locale-neutral results should explicitly state the desired language, region, and source.
The activation description is extremely broad, covering generic search, query, lookup, and research requests across many domains. This can cause the skill to trigger for a wide range of ordinary user intents, overriding more specific or safer skills and steering the agent into tool use and external data access when it may not be necessary.
The skill metadata emphasizes Chinese content and mandates use in certain contexts without confirming the user's language preference. This can bias routing and responses toward Chinese-language sources or workflows, reducing user control and potentially causing privacy, relevance, or policy issues when the user expected a different language or locale.
The file explicitly steers queries toward Chinese-language context and Chinese platforms for the doubao route even when the user has not opted into a Chinese-language or China-specific context. This can cause the agent to override user intent, bias source selection, and return region- or language-specific results that are less relevant, which is a prompt-safety and routing integrity issue rather than a code-execution flaw.
Line L03 states the resource is for '中文科技资讯', and the routing guidance later prioritizes 36kr for Chinese technology information. This is a natural-language locale preference presented as a default behavior, but the file does not offer opt-in/opt-out language choice or explain a region-specific requirement.
The file hard-codes language-based routing preferences, steering Chinese content to some sources and English content to others without requiring user preference or consent. This can bias source selection, reduce result relevance for multilingual users, and cause the agent to disclose queries to services the user did not choose.
The routing guidance uses broad everyday phrases like '找视频/影评/播客/文章' to trigger the skill without clear boundaries, which can cause over-activation and unintended delegation to external search tools. In an agentic system, vague triggers increase the chance that unrelated user requests are routed to networked search commands, expanding data exposure and reducing user control over tool use.
该文件整体以中文编写并将中文社区作为默认语境，但未说明是否允许用户选择其他语言，也未给出这是面向特定中文场景的明确限制。按照语言/locale 策略，若技能隐含强制特定语言而没有用户选择或正当说明，应视为自然语言政策问题。
The routing guidance is broad enough that a general request like '怎么看社交媒体' may trigger this skill without sufficient user intent, causing unnecessary access to social-platform-specific tooling or AI services. In an agent setting, over-broad dispatch increases the chance of collecting irrelevant third-party content, bypassing more appropriate search sources, or surprising users with platform-specific queries they did not explicitly request.
The file presents all user-facing guidance in Chinese and does not indicate that the skill is intended only for Chinese-speaking users or that another language option is available. This can violate language/locale policy when a skill effectively forces a specific language without user opt-in.
Line L24 explicitly describes Reddit as suitable for '英文社区问答', while the file overall presents guidance only in Chinese and does not mention any user opt-in or alternative language handling. This creates a mild language/locale policy concern because the skill guidance steers users toward a specific language context without documenting choice or consent.
No suspicious patterns detected.
