Security audit
Security checks for vulnerabilities and agentic risk
This skill is a simple RSS briefing workflow with disclosed, purpose-aligned web fetching and low-impact local history logging for deduplication.
Before installing, be aware that the skill will fetch RSS/article content from the configured sources and maintain a local pushed_history.log of pushed titles and links. Review or change the RSS list, and consider adding an explicit storage path and retention limit if you run it frequently.
The README description is written entirely in Chinese and does not indicate that users may choose another language or locale. Under the policy rule, forcing a specific language without user opt-in is a natural-language policy violation.
The file mandates Chinese headings and output template text such as 今日AI资讯, 标题：, 摘要：, and 链接： across the entire skill, but does not state that this is optional or region-specific. This creates a language-policy issue because the skill imposes a specific language/locale without user opt-in.
The skill explicitly instructs the agent to read and then append to a local pushed_history.log file, creating persistent side effects on the host without any user notice, confirmation, or scope limitation. This is dangerous because a skill can silently modify local state, affect future runs, and potentially overwrite or grow files in ways the user did not authorize, even if the stated purpose is benign deduplication.
No suspicious patterns detected.
