The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and execute the mcp-excalidraw-server package from the NPM registry using bunx at runtime.
[COMMAND_EXECUTION]: The skill uses shell commands to interact with the mcp-excalidraw-server CLI for rendering, exporting, and managing the Excalidraw canvas.
[DATA_EXFILTRATION]: The share command allows the agent to upload the diagram content to a public link service, which constitutes a network-based data transfer to an external endpoint.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied Mermaid diagrams and prompts to generate visual content, creating a surface for injection attacks.
Ingestion points: Processes external Mermaid code from .mmd files and natural language prompts from the conversation context.
Boundary markers: The instructions do not specify any delimiters or safety instructions to separate user data from agent instructions.
Capability inventory: Shell command execution via bunx, file system writes (PNG, SVG, and .excalidraw formats), and network uploads via the share command.
Sanitization: The instructions do not specify any validation or sanitization of the input data before it is processed by the rendering tool.

excalidraw-diagram