Security audit
Security checks for vulnerabilities and agentic risk
This Juejin automation skill needs Review because broad and inconsistent routing documentation is paired with authenticated account publishing and local file upload authority.
Review this before installing. Use it only for explicit Juejin tasks, and do not let it handle unrelated database, SQL, or generic automation requests. Treat publishing as account-affecting: confirm the Markdown file, destination account, tags, and whether it will remain a draft before allowing any upload or public publish. Be cautious about cookie storage because the artifact describes saving/loading session cookies but does not provide verifiable implementation details.
The description says to use the skill for broad scenarios like database operations, SQL queries, data storage management, efficiency improvement, automation, and workflow optimization. These triggers are much broader than the documented Juejin article ranking/publishing/downloading functions, making it unclear when the skill should or should not activate.
This skill documents outbound requests to an external Juejin API endpoint, which means user-supplied or locally sourced data may be transmitted off-host. In the context of a skill that can read local markdown and manage authenticated publishing via cookies, any external transmission increases the risk of unintended data disclosure if inputs are not tightly scoped and sanitized.
**API 接口**：
* 分类列表：`GET https://api.juejin.cn/tag_api/v1/query_category_briefs`
* 热门文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_all_feed`
* 分类文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_cate_feed`
* 标签列表：`POST https://api.juejin.cn/tag_api/v1/query_category_tags`
The documented POST request to the Juejin recommendation endpoint represents external data transmission to a third-party service. While expected for the feature, outbound connectivity is security-relevant because the skill also advertises broad automation and local file capabilities, creating a path for accidental disclosure if task boundaries are unclear.
**API 接口**：
* 分类列表：`GET https://api.juejin.cn/tag_api/v1/query_category_briefs`
* 热门文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_all_feed`
* 分类文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_cate_feed`
* 标签列表：`POST https://api.juejin.cn/tag_api/v1/query_category_tags`
The category feed endpoint is another documented external transmission surface. In this skill's context, external requests are legitimate, but they remain a true security concern because the skill can also process local content and authenticated sessions, so poor orchestration could leak more data than intended.
* 分类列表：`GET https://api.juejin.cn/tag_api/v1/query_category_briefs`
* 热门文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_all_feed`
* 分类文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_cate_feed`
* 标签列表：`POST https://api.juejin.cn/tag_api/v1/query_category_tags`
### 📝 功能二：文章自动发布
Querying category tags over the network is an external transmission channel. Although normal for the product, the skill's mixed capabilities make this more sensitive than a pure read-only client because users may not realize the same skill can both read local data and perform outbound requests.
* 分类列表：`GET https://api.juejin.cn/tag_api/v1/query_category_briefs`
* 热门文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_all_feed`
* 分类文章：`POST https://api.juejin.cn/recommend_api/v1/article/recommend_cate_feed`
* 标签列表：`POST https://api.juejin.cn/tag_api/v1/query_category_tags`
### 📝 功能二：文章自动发布
Creating article drafts via an external API is a true security-relevant action because it can transmit local markdown content, metadata, and potentially sensitive text to a third-party service. In this skill's context, the danger is higher than simple API reads because the feature couples local file access with authenticated outbound writes under a user session.
| 文章发布 | 通过掘金 API 创建草稿并发布，支持设置分类、标签、摘要、封面图 |
| 草稿管理 | 支持保存为草稿而不立即发布 |
* 创建草稿：`POST https://api.juejin.cn/content_api/v1/article_draft/create`
* 发布文章：`POST https://api.juejin.cn/content_api/v1/article/publish`
* 获取标签：`POST https://api.juejin.cn/tag_api/v1/query_category_tags`
Publishing an article is a state-changing external transmission that can send full content and metadata to Juejin under the user's authenticated cookie session. This is dangerous because an unintended or coerced invocation could publicly disclose sensitive local content, not just query remote data.
| 草稿管理 | 支持保存为草稿而不立即发布 |
* 创建草稿：`POST https://api.juejin.cn/content_api/v1/article_draft/create`
* 发布文章：`POST https://api.juejin.cn/content_api/v1/article/publish`
* 获取标签：`POST https://api.juejin.cn/tag_api/v1/query_category_tags`
**鉴权方式**：Cookie 鉴权（通过 Playwright 浏览器登录获取）
Fetching tags in the same authenticated publishing flow is security-relevant because it indicates the skill operates with a logged-in browser-derived cookie session. Any authenticated external request increases the consequence of prompt confusion or abuse, since it can act on behalf of the user account.
* 创建草稿：`POST https://api.juejin.cn/content_api/v1/article_draft/create`
* 发布文章：`POST https://api.juejin.cn/content_api/v1/article/publish`
* 获取标签：`POST https://api.juejin.cn/tag_api/v1/query_category_tags`
**鉴权方式**：Cookie 鉴权（通过 Playwright 浏览器登录获取）
Requesting article details from an external endpoint is a legitimate feature, but it still creates an outbound channel that may include user-selected URLs, identifiers, or follow-on content processing. Because the skill also supports saving data locally and downloading images, the combined workflow raises the chance of mishandling content or metadata.
| 图片处理 | 可选下载文章中的图片到本地 |
| 元数据保留 | 保留文章标题、作者、发布时间、标签等元信息 |
* 文章详情：`POST https://api.juejin.cn/content_api/v1/article/detail`
* 用户文章列表：`POST https://api.juejin.cn/content_api/v1/article/query_list`
## 快速上手
Querying a user's article list is an external transmission that can enumerate account-related content and metadata. In context, this becomes more sensitive because the skill supports bulk download and authenticated operations, increasing the risk of excessive collection or unintended scraping if triggers are too loose.
| 元数据保留 | 保留文章标题、作者、发布时间、标签等元信息 |
* 文章详情：`POST https://api.juejin.cn/content_api/v1/article/detail`
* 用户文章列表：`POST https://api.juejin.cn/content_api/v1/article/query_list`
## 快速上手
1. 确认运行环境满足依赖说明中的要求
The quick-start and operation-flow sections instruct users to 'call this skill' and 'choose the appropriate usage' but do not define specific trigger phrases, scope limits, or exclusion conditions. In a skill-routing context, this ambiguity can lead to unintended invocation for general user requests.
The file describes concrete Juejin capabilities such as fetching rankings, publishing Markdown articles, and downloading posts, but the documented inputs are generic review-style fields like content and strict_level. Those parameters do not align with the specific operations described elsewhere, indicating the documentation was copied from a different skill and now contradicts the stated behavior.
The surrounding documentation describes a Juejin skill for article ranking, publishing, and downloading, but the output example returns an audit-style grading report with fields like overall_grade, total_score, and security/compliance scores. This is not merely incomplete documentation; it actively describes a different kind of tool output than the Juejin functionality claimed elsewhere in the file.
No suspicious patterns detected.