The Agent Skills Directory

[SAFE]: No security threats were detected in the skill instructions or scripts. All components appear to be legitimate tools for SVG generation and rendering.
[COMMAND_EXECUTION]: The render.mjs script executes npm root -g to locate global Node.js packages and searches for system Chromium binaries in standard paths. These operations are limited to environment discovery for the rendering engine and do not involve untrusted input.
[DYNAMIC_EXECUTION]: The render.mjs script utilizes dynamic imports to load the playwright dependency from discovered paths, which is a standard approach for cross-environment compatibility in this context.
[INDIRECT_PROMPT_INJECTION]: The skill instructions define an attack surface where the agent processes user prompts to generate and execute JavaScript code for SVG creation. However, the skill explicitly includes a linter (lint.mjs) to verify that the generated outputs do not contain restricted elements like <script> or external references, significantly mitigating potential risks.

illustration-one-line