The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill uses npx to dynamically download and execute the excalidraw-cli package from the npm registry at runtime. This introduces a reliance on third-party code and external infrastructure (SKILL.md).\n- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands where complex JSON data, generated from user input, is passed as a command-line argument (e.g., npx -y excalidraw-cli@0.0.2 create --json '...'). This pattern is vulnerable to command injection if the input contains shell metacharacters that are not properly escaped by the execution environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user requirements to generate diagram content and shell commands, creating an attack surface for malicious input.\n
Ingestion points: User prompts and requirements provided to the /excalidraw-diagram command (SKILL.md).\n
Boundary markers: There are no explicit delimiters or instructions to treat user-provided data as potentially untrusted.\n
Capability inventory: Execution of shell commands and file system write access (references/cli-usage.md).\n
Sanitization: The skill lacks explicit sanitization or validation steps for the generated JSON before it is used in a shell command.\n- [EXTERNAL_DOWNLOADS]: The skill fetches the excalidraw-cli tool from the npm registry and provides functionality to upload data to excalidraw.com. These external interactions involve third-party data transit and code execution.

excalidraw-diagram