The Agent Skills Directory

[SAFE]: The skill operates within expected security parameters for its stated purpose. Its use of system utilities and external font CDNs is transparent and limited to the local development environment.
[COMMAND_EXECUTION]: The workflow includes steps to open the generated HTML brand kit using standard platform commands (open on macOS, xdg-open on Linux, and start on Windows). These are benign commands used solely for user review of the output files.
[EXTERNAL_DOWNLOADS]: The generated HTML files include <link> tags to fetch typefaces from the Google Fonts CDN. Additionally, the documentation mentions installing the skill monorepo via npx. These interactions involve well-known, trusted services and vendor-controlled resources.
[PROMPT_INJECTION]: The skill ingests untrusted external data when reading existing branding HTML or landing pages to extract design tokens. This represents a potential indirect prompt injection surface. \n
Ingestion points: Step 1 involves the agent reading content from user-provided branding files or URLs. \n
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' commands for the content being processed. \n
Capability inventory: The skill possesses the ability to write to the local file system (creating a themes/ directory) and execute system 'open' commands. \n
Sanitization: No explicit sanitization, validation, or filtering of the content extracted from external files is described in the workflow.

brand-kit