Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only Chinese social-media content planning skill whose broad routing is visible and mostly limited to advice and generated content, not account control or system access.
Install if you want a Chinese-language assistant for social-media planning and drafts. Be aware it may activate on broad creator-related requests, and avoid sharing private backend screenshots, contracts, or business metrics unless you are comfortable using them as input for analysis.
The skill advertises very broad natural-language activation tied to generic requests, making accidental invocation likely. Over-broad triggering can cause the skill to intercept unrelated user intent, leading to prompt-scope confusion, unintended routing, and reduced user control over which capability is active.
The guidance explicitly says users can invoke the skill with unrestricted natural language and that the system will infer missing details, which encourages trigger-on-ambiguity behavior. In a multi-skill or agentic environment, this increases the risk of unintended execution paths and makes it harder to enforce least surprise and explicit consent.
The routing rules use catch-all triggers like "自媒体" and direct semantic matching on broad requests, which can cause the skill to activate on a wide range of normal conversation. Because this skill orchestrates many submodules, unintended activation can cascade into extensive content-generation workflows, amplifying scope hijack and misrouting risks.
The README advertises very broad natural-language triggers such as '帮我想几个[领域]的选题' and '帮我做一篇完整内容', which can overlap with ordinary user requests and cause the skill to activate unintentionally. In a multi-skill environment, this increases prompt-routing ambiguity and could lead to unexpected execution of a large, chained workflow spanning multiple modules.
The manifest description is entirely in Chinese and specifies Chinese trigger phrases as the expected invocation format, with no indication that users may choose another language. This can constitute a language-policy issue because it imposes a specific language behavior without explicit user opt-in or documented regional justification.
The trigger instructions require Chinese keywords and examples only, such as "一人MCN" and other Chinese natural-language requests, while the document does not offer multilingual alternatives or an opt-in choice. This presents the skill as enforcing a single language rather than adapting to user preference.
The synonym list is intentionally non-exhaustive and includes commonplace phrases such as asking what to write or how to name something. While useful for UX, this creates ambiguous matching conditions that can overreach into normal assistant interactions and trigger the skill when the user did not intend to use it.
The natural-language instructions and usage examples are all presented in Chinese, with no indication that users may choose another language or that the skill is intentionally limited to Chinese-speaking users. This can be a language-policy concern when a skill effectively forces one language without explicit opt-in or justification.
No suspicious patterns detected.
