The Agent Skills Directory

[SAFE]: The skill uses a conversational interview pattern to collect user input, which is a standard and safe method for interactive document generation. It utilizes platform-native tools like ask_user_input for structured data collection.
[SAFE]: File system operations are confined to the designated output directory (/mnt/user-data/outputs/) and use standard presentation tools (present_files) to share the result with the user, adhering to best practices for data handling.
[SAFE]: No network requests, external code downloads, or third-party dependencies are present. The skill operates entirely within the local agent environment using provided templates.
[SAFE]: The logic for generating kebab-case identifiers and Markdown assembly is transparent and does not involve dynamic execution of user-provided content.

rfc