Security audit
Security checks for vulnerabilities and agentic risk
This skill clearly aims to publish Xiaohongshu posts, but it can automatically click the final publish button on a live logged-in account without a clear confirmation gate.
Install only if you intend the agent to operate a real logged-in Xiaohongshu creator account. Before use, verify the exact title and content yourself and avoid broad publish requests unless the skill is updated to require an explicit final confirmation before clicking publish.
This markdown file contains user-facing instructions only in Chinese, including headings and command explanations. The policy allows locale constraints only when users are given a choice or when the restriction is clearly documented and justified, neither of which appears here.
The README describes automated end-to-end posting and 'one-click publish' on a real Xiaohongshu creator account, but does not clearly warn users that the action results in an actual public post. This can mislead users into triggering irreversible or externally visible actions without informed consent, especially in an automation context where the skill may execute quickly and with little friction.
The skill automates posting to a live Xiaohongshu creator account and explicitly notes that drafts may be auto-saved, but it does not present a clear warning that execution can cause real external side effects. This increases the risk of accidental publication, unintended draft creation, or reputational harm if a user invokes the skill without understanding it acts on a real logged-in account.
The usage guide describes a workflow that culminates in clicking a final "publish" button and explicitly says it will complete publication, but it does not clearly warn that the skill performs an externally visible, irreversible action on the user's behalf. In an agent setting, missing or weak disclosure around real-world side effects increases the risk of unintended public posting, especially if invoked from a broad natural-language request.
The workflow performs an irreversible publish action automatically by calling the publish step immediately after navigation and form-filling, with no interactive confirmation, preview gate, or explicit user consent check at the final step. In an automation skill that can post public content, this creates a real risk of unintended publication, reputational damage, or accidental disclosure if the title/content is wrong or the page state is unexpected.
All user-facing natural-language instructions in the file are Chinese, and the document does not indicate that the skill is region-specific by policy or that users may opt into another language. Under the language/locale policy, forcing a specific language without user opt-in can be a natural-language policy violation.
The natural-language instructions and headings are entirely in Chinese, and the file does not indicate that language selection is optional or user-configurable. Under the stated policy, forcing a specific language without user opt-in can be a locale/language policy issue unless clearly documented and justified.
自然语言内容全部以中文呈现，且未说明这是仅面向中文用户的区域性技能，也未提供其他语言选项或用户 opt-in。根据规则，若技能强制特定语言而无明确选择机制或合理限定，属于潜在的语言/locale 政策问题。
The example trigger phrases for direct agent use are broad conversational requests such as asking to publish content, which could overlap with normal discussion and make accidental invocation more likely. Because this skill performs account actions and public posting, ambiguous activation language is riskier than for read-only or informational skills.
The file presents the skill primarily in Chinese and all runtime status messages are Chinese, but it does not state that the skill is region-specific or provide any opt-in or alternative language behavior. That can violate language/locale policy where users must be given a choice unless the locale restriction is explicitly justified.
