The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (MEDIUM): The skill requires an external CLI tool @swiftlysingh/excalidraw-cli which is not from a pre-approved trusted source. It also suggests using npx to execute this package directly from the registry without version pinning, which poses a risk of supply chain attacks if the package is compromised.
[COMMAND_EXECUTION] (LOW): The skill generates and executes shell commands using npx or npm. While these are based on user-provided descriptions, the structure uses heredocs and inline strings which could be manipulated if the agent doesn't properly sanitize the generated DSL before execution.
[INDIRECT_PROMPT_INJECTION] (MEDIUM): Category 8 analysis:
Ingestion points: Processes natural language descriptions from the user to generate diagrams (SKILL.md).
Boundary markers: None. The DSL is interpolated directly into shell commands (--inline "YOUR_DSL_HERE").
Capability inventory: Execution of external CLI via npx and file writing to the local filesystem (-o flowchart.excalidraw).
Sanitization: No explicit sanitization of the DSL input is mentioned before it is passed to the shell command. An attacker could potentially craft a diagram description that attempts to break out of the DSL string and execute arbitrary shell commands.

Excalidraw Flowchart