The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions suggest a validation step using a local Node.js script (node <walkthrough-record>/scripts/record.mjs scenario.yaml --validate-only). This is a standard development workflow for linting and schema validation of the generated YAML and is not considered malicious.
[DATA_EXPOSURE]: The documentation includes proactive security advice for handling sensitive data. It guides the user to identify sensitive UI elements and add them to a mask list to prevent them from being recorded. Additionally, it explicitly warns that storageState files (which contain session cookies/authentication tokens) must be kept out of version control systems like git, demonstrating safe secret management practices.

walkthrough-storyboard