The Agent Skills Directory

Instructional Content: The skill consists entirely of natural language instructions and formatting rules for an AI agent to follow when creating Mermaid diagrams. It defines best practices for visual documentation and syntax safety to ensure diagrams render correctly.
Indirect Data Ingestion: The skill instructs the agent to ground diagrams in 'inspected source' code. While the agent processes source code (which is technically external data), this is a standard documentation task. The risk of indirect prompt injection is minimal as the output is restricted to Mermaid diagram syntax within Markdown fences.
Syntax Validation: The instructions include specific guidance on avoiding reserved words and illegal characters (like semicolons or unescaped angle brackets) to prevent parsing failures in the rendering engine.

mermaid-diagrams