The Agent Skills Directory

[COMMAND_EXECUTION]: The skill directs the agent to execute shell commands using the mmdc (Mermaid CLI) tool. The command mmdc -i <input.md> -o /tmp/<name>-rendered.md -a /tmp/<name>-mermaid-artifacts involves variable interpolation for filenames and paths. If the agent does not properly sanitize these inputs, it could lead to command injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests untrusted data from markdown files and uses tool output to drive subsequent actions.
Ingestion points: The agent reads the content of input markdown files and parses raw stderr and stdout from the mmdc command (SKILL.md, multi-diagram-validation.md).
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious instructions embedded within the diagrams or error messages.
Capability inventory: The skill is capable of executing shell commands (mmdc) and reading/writing files in the /tmp directory (SKILL.md).
Sanitization: The instructions do not define any sanitization or validation logic for the external content before it is processed by the agent.

create-mermaid-diagrams