Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed Markdown-only web search skill that sends user search terms to listed third-party search and content sites.
Install only if you are comfortable with your search terms being sent to the selected external search engines and content sites. Avoid searching confidential, personal, legal, medical, or sensitive business information through broad third-party engines.
The skill name, description, and usage examples are entirely Chinese, and the examples imply the skill is invoked in Chinese only. There is no statement offering language choice or explaining that the skill is intentionally limited to a Chinese-language or region-specific context.
The example trigger phrase is broad enough to overlap with ordinary user speech, which can cause the skill to activate unintentionally during unrelated conversations. In an agent environment, unintended activation may route user queries to external web-scraping behavior without clear user intent, increasing the risk of privacy leakage, prompt hijacking exposure from fetched pages, or surprising actions.
The skill explicitly routes user queries to many third-party search and content sites via web_fetch, but it does not warn users that their search terms will be transmitted to external services. Search terms can contain sensitive business, financial, health, or personal information, and broadcasting them to dozens of sites materially increases privacy and data-leakage risk.
The description explicitly presents the skill as '中文网页搜索' and the document is written entirely in Chinese, but it does not state that this locale restriction is intentional or optional for users. The policy requires that language or locale constraints be either user-selectable or clearly justified.
The package description is written entirely in Chinese and presents the skill as a Chinese web search capability without indicating any language choice or opt-in. Under the policy, forcing a specific language/locale can be a natural-language policy violation unless the constraint is explicitly documented and justified.
No suspicious patterns detected.
