Security audit
Security checks for vulnerabilities and agentic risk
This is a no-code automation planning guide with no hidden execution or persistence, though its declared exec capability is broader than the content needs.
Installers should treat this as a planning and checklist skill. Keep direct control over any connected Zapier/Make/n8n, CRM, billing, email, or social-media actions, and be cautious about granting exec where a Markdown guide does not need local command execution.
The trigger keywords are broad, common terms related to automation tools and maintenance, which raises the chance of accidental invocation in unrelated conversations. Unintended activation becomes more concerning here because the skill also declares exec capability, potentially exposing users to actions they did not explicitly request.
The skill is primarily advisory/documentation content about no-code workflow design, yet its manifest exposes the exec capability. That expands the agent's authority beyond the stated purpose and creates unnecessary risk of shell or system command execution if the skill is invoked or later extended unsafely.
The skill advertises MD+EXEC and discusses platform-assisted execution/configuration, but it does not clearly warn users that system operations or external integrations may be performed. This can undermine informed consent and increase the risk of surprising or unsafe actions, especially when connected SaaS tools or local execution are available.
文件从 displayName 到主体说明均固定为中文，未向用户提供语言/locale 选择，也未说明该技能仅面向特定中文场景或地区。根据语言/locale 政策，若技能强制特定语言而无用户选择，属于自然语言层面的策略问题。
No suspicious patterns detected.
