Security audit
Security checks for vulnerabilities and agentic risk
This is a Chinese blog rewriting and cover-generation skill whose network and file-writing behavior is disclosed and aligned with its stated purpose.
Install this if you want a Chinese-language article rewriting workflow. Do not give it private, paywalled, or logged-in URLs unless you are comfortable with the agent opening and processing that content; cover generation will create a local HTML file in the workspace.
The manifest description and main heading/content present the skill as operating in Chinese and define style rules in Chinese without any opt-in or alternative language option. Under the policy, language constraints should either be optional for the user or explicitly justified as region-specific.
The document is entirely framed as prescriptive guidance for writing style in Chinese, including specific pronoun and phrase recommendations, but it does not state that the skill is limited to Chinese-language content or offer users a language/locale choice. Under the policy, forcing a specific language or locale without opt-in is a natural-language policy violation.
The skill explicitly tells the agent to fetch user-provided links with web_fetch or browser, but it does not disclose that this causes external network access and transmits the supplied URL to third-party services or remote sites. In a rewriting skill, users may provide private, unpublished, or tracking URLs; silently fetching them can leak sensitive information and create privacy or compliance issues.
This markdown file explicitly instructs use of the write tool to save an HTML file locally, which affects the user's filesystem. The description states the behavior, but it does not present it as a caution or warning about creating a local file or where it will be stored.
The document declares lang="zh-CN", and the visible title/instructions are also in Chinese, which indicates the template is fixed to a specific language/locale. Under the policy criteria, forcing a specific language without opt-in or justification is a natural-language policy concern.
The document declares lang="zh-CN", which fixes the content to a specific language/locale. Under the policy, locale constraints should either be optional for users or clearly justified as region-specific; this file provides no such opt-in or justification.
The sample HTML explicitly sets lang="zh-CN", and the surrounding content is entirely Chinese-oriented, but the file does not state that this template is intentionally limited to Chinese-language use cases or offer alternatives. Under the policy criteria, forcing a specific language or locale without opt-in can be a natural-language policy violation.
This template repeats the same locale constraint by setting lang="zh-CN" while providing no explanation that the template is region- or language-specific. Because the file gives reusable templates, the fixed locale may propagate to users who did not intend a Chinese-language document.
The locale is again hard-coded in a reusable example, and the file does not indicate that users should adjust it based on their own language needs. That creates the same language-policy concern across multiple templates.
The guideline instructs the agent to use write to create an HTML cover file in the workspace without warning the user that a file will be created and stored locally in the session workspace. While lower risk than network exfiltration, undisclosed file creation can surprise users, leave behind generated artifacts, and be problematic if the content contains sensitive or proprietary material.
