Security audit
Security checks for vulnerabilities and agentic risk
This is a read-only analytics dashboard skill with some documentation and scoping caveats, but no artifact-backed malicious behavior.
Install only where dashboard-mcp is configured for authorized users, because responses may include operational, tenant, event, queue, and agent-status data. Treat the Cookie-management line as a documentation defect and prefer narrowing read permissions to dashboard-mcp-only access if the platform supports it.
声明描述的是一个面向数据看板查询的 Skill，应具备获取实时统计、系统健康等看板数据的能力。但代码片段实际上只是测试脚本：读取本地 SKILL.md 文件并断言其文档结构完整，没有任何网络请求、dashboard-mcp 访问、实时统计查询、系统健康检查或租户/队列/订阅数据获取逻辑。因此该代码的主要目的与声明用途显著不一致，属于明显描述-行为不匹配。
The skill exposes a read capability via the manifest but does not define a clear allowlist or permission boundary for what may be read. In a dashboard skill that can surface operational and tenant data, missing tool scoping increases the chance of unintended file or data access if the implementation later expands or if the runtime interprets generic read access broadly.
The trigger list includes broad phrases such as morning-briefing or agent-status style requests without strong constraints, which can cause the skill to activate in loosely related contexts. For a skill that surfaces system health, tenant overviews, and event feeds, overbroad activation can expose operational or tenant-specific data when the user did not clearly intend a sensitive dashboard query.
L021 写明“本技能提供/Cookie管理等能力”，但 manifest 的 L003 明确写“不触发:...Cookie管理”，后续工作流 L040-L068 也全部是 dashboard-mcp 的只读看板查询。这属于文档文字与技能真实意图/实现范围直接冲突，而非单纯信息不完整。
The skill handles potentially sensitive operational and tenant-specific dashboard data but does not warn users about that sensitivity or expected access controls. Without a disclosure boundary, users may request or receive system-health, event-stream, or tenant-overview information without understanding the confidentiality implications.
The manifest description is written entirely in Chinese and embeds Chinese trigger phrases, which indicates a language-specific interaction model. The file does not state that the skill is region-specific or provide an opt-in or alternative language path.
This Python file contains multiple natural-language comments and docstrings in Chinese, including the primary file description and test descriptions, without indicating that Chinese is optional or required for a justified region-specific purpose. Under the policy rule for language/locale, this can be considered a forced language choice in user-facing or maintainer-facing natural-language content.
No suspicious patterns detected.
