Security audit
Security checks for vulnerabilities and agentic risk
The skill is for chart generation, but it asks for command execution and file writes while the referenced local script is missing, leaving the actual command target unclear.
Review this carefully before installing. It may cause an agent to execute a local chart.sh command and write chart files, but the package does not include the script it says it will use. Install only if you can verify the intended chart.sh executable and direct outputs to a safe path.
The skill describes saving HTML/SVG outputs but does not clearly warn users that local files will be written, where they will be written, or that existing files may be overwritten. In an agent setting, undisclosed write behavior can surprise users, alter working directories, or clobber files when combined with automatic execution.
The trigger-scope text is overly broad and maps many generic phrases to the skill, increasing the chance that an agent invokes this exec-capable skill in contexts where the user did not explicitly request chart generation. Because the skill advertises exec and file-producing behavior, over-triggering can lead to unintended local command execution or artifact creation.
The skill declares itself as MD+EXEC and lists exec, but it does not prominently warn that using it may invoke a local Bash script. In agent environments, undisclosed local command execution materially raises risk because users may believe the skill is purely descriptive or markdown-based when it can actually run shell commands.
L233-L244 表明原生 PowerShell 不支持，要求使用 Git Bash/WSL，并整体围绕 Bash 环境设计。虽然这更接近平台约束而非严格语言政策，但它对特定本地环境进行了硬性限定，且未为偏好其他本地化终端环境的用户提供可选路径或明确的适用范围说明。
L262-L269将免费版定位为受限版本，并明确列出“❌ 多图表组合仪表盘的交互式联动”为不支持能力；但同一文件前文 L057、L162、L237 又把 dashboard 作为可用命令和推荐用法。这种文档层面的表述容易让人误解免费版到底是否支持仪表盘相关能力，属于意图说明与功能说明不一致。
