The Agent Skills Directory

[SAFE]: The skill's primary function is to perform a 'read-only conversion audit.' The instructions explicitly forbid page writes, generation, or publishing, stating that the skill 'changes nothing.'
[SAFE]: All external URLs referenced in the documentation point to legitimate industry authorities (e.g., Nielsen Norman Group, Baymard Institute), regulatory bodies (FTC, ASA, ASCI), and established technology platforms (Shopify, Google Merchant Center).
[SAFE]: The skill uses a managed toolset (MCP) for gathering analytics, catalog data, and page source. There is no evidence of hardcoded credentials or sensitive file access.
[SAFE]: The repository includes detailed guidelines for 'Managed Motion' and 'Quiz Authoring' which describe a high-security sandbox environment. These guidelines mandate the use of a restricted DSL and a compiler that automatically rejects dangerous JavaScript APIs (like window.location, fetch, or eval), global DOM access, and raw timers.
[SAFE]: No obfuscated content, dynamic context injection, or persistence mechanisms were detected. The skill is composed of pure markdown instructions and structural configurations.

audit