Security audit
Security checks for vulnerabilities and agentic risk
The skill mostly matches its content-conversion purpose, but it also describes automated publishing through logged-in creator accounts without clear confirmation or safeguards.
Install only if you want a Chinese-market social-content drafting workflow. Treat it as safe for local draft generation, but do not allow browser automation, uploads, or publishing from logged-in accounts unless you explicitly review each post and confirm each platform action.
Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.
# social-content-converter
> 一鱼三吃内容改写器 — 输入一篇文章，自动改写成抖音、小红书、B站三个平台的专属版本
The skill description is written as a direct Chinese-only user-facing instruction and presents the tool as accepting input and producing outputs in Chinese, with no indication that other languages are supported or that Chinese is an optional mode. This is a natural-language locale policy concern because it imposes a specific language without user opt-in or documented justification.
The skill advertises behaviors that read from and write to local files, including batch processing and generating output directories, but it does not declare any explicit tool scope or permission boundaries. This creates ambiguity about what filesystem access is intended and can lead to over-privileged execution, unexpected file modification, or abuse if the agent/runtime grants broader access than the user realizes.
The trigger phrases are broad conversational terms like '改写内容' and '生成各平台文案', which can match many ordinary user requests and cause the skill to activate unexpectedly. Over-broad activation is risky here because the skill can perform file-producing workflows and potentially steer users into automated publishing flows they did not explicitly request.
The description defines output specifically for Chinese platforms and is written as if the generated content will be tailored in that locale by default, without stating that users may choose another language or locale. This can violate language/locale choice expectations when the skill is invoked in a broader multilingual environment.
The skill describes automated publishing to logged-in creator platforms and generation of output files without any visible safety gates, confirmation steps, or warnings about account-impacting actions. In context, this is more dangerous because browser automation against authenticated sessions can post content, upload media, or modify creator accounts at scale if triggered unintentionally or with malicious input.
This code file contains natural-language descriptions and CLI output entirely in Chinese, including the top-level docstring and argument/help text, which establishes a fixed language experience. Under the policy, forcing a specific language without user opt-in is a locale/language policy violation unless the constraint is clearly documented and justified.
The skill is described as converting content into platform-specific versions, and the CLI advertises a --platforms option for selecting targets. However, the code ignores args.platforms and convert_all() always processes douyin, xhs, and bilibili, so the actual behavior is broader than the user-facing interface suggests.