The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill dynamically generates a Python script at /tmp/serve_<client>.py using SimpleHTTPRequestHandler to provide a local preview of the generated design system. This involves runtime code generation and execution to serve files on a local port.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the ingestion of untrusted data from external websites.
Ingestion points: Step 1 in SKILL.md fetches content and CSS from user-provided URLs via WebFetch and curl.
Boundary markers: No explicit delimiters or instructions are provided to the agent to isolate or ignore potentially malicious content within the fetched data.
Capability inventory: The skill can execute shell commands, write files to the workspace, and perform deployments to Netlify.
Sanitization: While it filters CSS colors using regex, the narrative text used for brand grounding is processed without sanitization.
[COMMAND_EXECUTION]: The skill executes multiple shell commands for web scraping and data processing, including curl, grep, and sort. It also uses npx to run deployment tools.
[EXTERNAL_DOWNLOADS]: The skill retrieves raw HTML and CSS from arbitrary external URLs and uses npx to download and execute the netlify-cli tool for production deployments.

design-system-generator