Security audit
Security checks for vulnerabilities and agentic risk
The skill is coherent for interview preparation, but it can create GitHub Issues and run external search/scraping workflows from broad triggers without clearly requiring confirmation or a chosen target repository.
Review this carefully before installing if your agent has GitHub write access. Use it only with an intended repository selected, and confirm before letting it create or manage Issues; otherwise treat it as suitable for read-only planning and mock-interview guidance.
The trigger list includes generic terms such as '面试', '技术面试', and '面试准备', which are broad enough to match normal conversation and could activate the skill unintentionally. Because the skill then orchestrates external search/fetch operations and may create GitHub Issues, accidental invocation can lead to unanticipated outbound actions and repository-side changes.
The skill workflow normally creates GitHub Issues, but the description and usage flow do not prominently warn users that repository modifications are a standard side effect. This undermines informed consent and increases the chance that a user triggers the skill expecting analysis only, while the agent performs write actions against a connected GitHub repository.
The trigger "准备{公司}{岗位}面试" is broad enough to match ordinary user conversation about interview preparation, which can cause the workflow to activate unintentionally. In a skill that performs external search, scraping, issue creation, and content generation, accidental activation can lead to unintended network actions or repository modifications without sufficiently explicit user intent.
The trigger "制定本周刷题计划" is a generic planning phrase that overlaps with normal assistant usage, so the skill may activate when the user only wants advice rather than workflow execution. Because this entry point chains search, scraping, issue creation, and card rendering, an ambiguous trigger increases the risk of unintended side effects and surprise actions.
This README presents the skill primarily in Chinese and does not state whether users can opt into another language or locale. Under the language/locale policy, forcing a specific language without user opt-in can be a natural-language policy concern when the skill description gives no alternative.
The trigger "Review 我的{答案}" is less precise than a dedicated command and can collide with normal requests for feedback, particularly in a mixed-language conversation. Although this entry point only invokes the review skill, unintentional activation could still cause the agent to route content into a specialized evaluation workflow the user did not explicitly request.
No suspicious patterns detected.
