The Agent Skills Directory

COMMAND_EXECUTION (LOW): The skill documentation facilitates the use of the mmdc command-line tool. Running this tool requires the agent to execute shell commands with user-defined input/output file paths, which could be exploited for command injection if not properly sanitized.\n- REMOTE_CODE_EXECUTION (LOW): Documentation in mermaid-cli/already-installed-chromium.md explains how to set the executablePath for Puppeteer. This setting allows the execution of an arbitrary binary, which is a potential risk if the configuration file is controlled by an attacker.\n- PROMPT_INJECTION (LOW): Mermaid diagram syntax allows for user-defined text in nodes and labels. This acts as a surface for indirect prompt injection, where an attacker could embed instructions intended for the AI agent.\n- EXTERNAL_DOWNLOADS (LOW): The skill recommends installing the @mermaid-js/mermaid-cli package. As the mermaid-js organization is not on the predefined list of trusted entities, this represents an unverifiable external dependency.

mermaid