Security audit
Security checks for vulnerabilities and agentic risk
This skill appears to be a skill-writing helper, but its trigger text is inconsistent and broad enough that a write/exec-capable skill could be invoked for unrelated content-writing requests.
Review this before installing. It does not show malware, exfiltration, or persistence, but its description should be narrowed to skill creation only and should clearly state when file writes or shell commands may be used. Avoid installing it if you expect it to handle general marketing or writing tasks without local file or command authority.
The manifest/use guidance says the skill is for marketing and content writing, while the body actually documents a skill-construction tool. This high-severity semantic mismatch can cause incorrect routing by the agent, leading it to apply exec/write-enabled instructions in inappropriate contexts where the user only asked for text generation.
The manifest frames the skill as a lightweight skill-authoring tool, but its description also advertises broad marketing/content-generation use cases. This mismatch can cause the agent to invoke the skill in contexts unrelated to safe skill authoring, increasing the chance of unintended use and capability overreach.
The description uses broad trigger phrases such as marketing copy, writing content, title optimization, and content creation, which are common across many benign requests. Overly broad invocation criteria increase the risk that the agent selects this higher-privilege skill when a narrower, safer skill would suffice.
The skill declares read, exec, and write capabilities, but the document does not prominently warn users about command execution, file modification, or their side effects. In combination with the skill's ambiguous scope, this increases the risk of unexpected local command execution or file changes under an imprecise invocation path.
The document describes generic create/query/modify/delete/export/import operation modes and structured parameter handling that are much broader than the stated role of generating structured skill files. In a skill with write and exec capabilities, such generic operational language can justify actions beyond user expectations and widen the effective authority of the skill.
The example description says to use the skill when the user mentions broad concepts like formatting or code style, which are ambiguous and could match many unrelated tools. While presented as an example, it reinforces imprecise invocation patterns that can degrade routing safety.
The skill name and description are presented in Chinese with embedded English phrases like 'Use when,' and the document consistently models this mixed-language output as the expected format. There is no statement that the user may choose their preferred language or locale, which can conflict with language-choice policies.
No suspicious patterns detected.
