The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute shell commands. It specifically calls the drawio CLI for exporting diagrams (e.g., drawio -x -f <format> -e -b 10 -o <output> <input.drawio>) and uses platform-specific commands like open, xdg-open, and cmd.exe /c start to open the generated files.
[EXTERNAL_DOWNLOADS]: The skill utilizes the WebFetch tool to download an XML reference guide from the official Draw.io repository at https://raw.githubusercontent.com/jgraph/drawio-mcp/main/shared/xml-reference.md.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because user-controlled data is interpolated into shell commands without explicit sanitization.
Ingestion points: The skill derives filenames and diagram content from user requests (e.g., 'png flowchart for login' results in a filename including 'login').
Boundary markers: There are no explicit instructions or delimiters used to ensure the agent sanitizes or ignores potential shell metacharacters in user-supplied filenames.
Capability inventory: The agent has access to Bash for command execution and Write for file creation.
Sanitization: No sanitization logic is defined for processing user input before it is used in the Bash command templates.

drawio