The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing local Node.js scripts (tools/generate-image.mjs, tools/process-sprites.mjs) and standard Unix utilities (grep, cut, rm, sips) to perform its primary functions. These tools are used for secret retrieval, asset generation, and file cleanup.
[INDIRECT_PROMPT_INJECTION]: The asset generation workflow is vulnerable to command injection because user-supplied prompts are interpolated directly into shell commands: node tools/generate-image.mjs "<prompt>".
Ingestion points: The <prompt> variable in SKILL.md is populated by user requests such as "create a sprite" or "make game art".
Boundary markers: The input is enclosed in double quotes within the instructions, but this provides minimal protection against shell escaping or subshell execution.
Capability inventory: The environment allows for file deletion (rm), image manipulation via sips, and execution of local JavaScript files via node.
Sanitization: No sanitization or validation steps are defined in the instructions to escape shell metacharacters (e.g., ;, &, |, `) before the prompt is passed to the shell.

image-generation